ProgrammableMarketing/simantic.com (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 7
  • This quarter · 11
  • Big bets · 3
  • Wildcards · 2
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. Simantic
Simantic logo

growth playbook · Nov 10, 2026

simantic.com (opens in a new tab)

Simantic growth playbook

Diagnosis

Simantic is a four-person Y Combinator Fall 2026 company from Waterloo that launched on October 29, 2026 with a from-scratch Rust firmware simulator, a CLI, an API, an MCP server and a pull-request reviewer that actually runs the code. The product is deeper than the website. The site never says which chips it runs.

The Cloud Sim page is one sentence. Pricing is in dollars of credit with no explanation of what they buy. The homepage marks the CLI and Python as "coming soon" when both exist. There is a CI action on GitHub that the site never mentions, the validation page has no results and the blog is empty.

This playbook first makes the product legible (a public chip list, a real quickstart, credit math, CI and Python surfaced), then builds proof (validation results, real bugs found in public firmware, a weekly model changelog), then opens the larger bets: board-level electronics simulation from the new SPICE hire, chip-maker partnerships and a local Pyrite licence for regulated teams.

Growth Score

6/10

Biggest Opportunity

Embedded engineers search by part number and buy on proof. Simantic can win those searches with a public page per supported chip, published silicon-versus-simulation results and real bugs caught in known open-source firmware.

Biggest Risk

Renode (open source, 1200+ boards), Wokwi (free browser simulation and CI minutes) and hardware-first agents like Embedder already have the audience. A four-person team that spreads across too many chips, interfaces and channels before proving accuracy could be dismissed as "another simulator".

01

Start here

The three moves to make first. Each one is written up in full below.

  1. 1

    Publish one public list of supported microcontrollers, and link it from the homepage, Cloud Sim and pricing

    Export `sim models --show-unmodeled` to a JSON file and render it as a table at /mcus.

    High impact First result in 2 weeks
  2. 2

    Turn the Cloud Sim page into a working quickstart: one install, one run, real output

    Copy the docs quickstart and one real run's output onto /cloud-sim.

    High impact First result in 2 weeks
  3. 3

    Show what $10 and $200 of credits actually buy, and label Community as free

    Pull last month's median cost per review and per simulated minute from billing.

    High impact First result in 2 weeks

02

Your idea portfolio

23 ideas, grouped by when to run them. Open any idea for how to run it: the steps, who to approach, copy you can use, and how to measure it.

  • 8 ideas are marked Unexpected: plays you are unlikely to have tried.

Every idea placed by the effort it takes against the impact it could have. Quick wins sit top left; select a number to open that idea. The filters below narrow the list by status, timing, category and what Programmable can build.

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 5Big bets · 7Fill-ins · 9Reconsider · 2
Effort →

Showing 23 of 23 ideas

This week 7

Quick wins with a clear first action.

  • 01

    An embedded engineer's first question is "does it run my chip?", and simantic.com never answers it where they decide. The homepage, the /cloud-sim page and the pricing page name no supported microcontroller (the only chip on the homepage is the STM32F072RB in a sample review). The lists that exist are buried in the docs and differ in shape: /docs shows RP2040, STM32F072RB, STM32F401RE and "..." plus an STM32H7 status note, the guide names G474, F767, nRF52840 and ESP32 parts as examples, and launch coverage lists STM32F0/F4/H7, ESP32-S3/C3 and RP2040 with "new models" weekly. The full answer is only in `sim models`, after installing the CLI. And a pull request for an unsupported chip gets a static review that is "never approved", so a visitor who guesses wrong finds out after connecting GitHub.

    High impactLow effort First result in 2 weeks Founder/CTO

    First action

    Export `sim models --show-unmodeled` to a JSON file and render it as a table at /mcus.

    How to measure it

    Signup rate after a /mcus view

    Now
    –
    Target
    2x the site-wide signup rate
    By
    60 days

    Steps

    1. 1

      Export the list

      Dump `sim models --show-unmodeled` with status, peripherals and first version to JSON.

      Founder/CTO 1 day

    2. 2

      Build /mcus

      A filterable table by vendor, family and status, rendered at build time in the Astro site.

      Founder/CTO 2 days

    3. 3

      Write chip pages

      One page per released chip: example command, sample output, buses modeled, PR-review setup, known simplifications.

      Founder/CEO 1 week

    4. 4

      Add chip requests

      A short form for part number, use case and team size; requests feed the modeling backlog.

      Founder/CEO 1 day

    5. 5

      Link it everywhere

      Homepage hero line, /cloud-sim, /pricing, /docs/pr-review and the static-review message from the bot.

      Founder/CTO 2 hours

    6. 6

      Automate updates

      Regenerate the page on every sim release so the list never drifts from the binary.

      Founder/CTO 2 days

    7. 7

      Measure

      Signups after /mcus views, chip-page search impressions and top requested parts.

      Founder/CEO monthly

    First result
    2 weeks
    Cost
    About a week of founder time; no new tools.
    Owner
    Founder/CTO

    Smallest test

    Add one line under the homepage hero, "Runs today on STM32F0/F4/H7, ESP32-S3/C3 and RP2040", linked to the docs list, and compare signup rate for two weeks.

    Why it works

    Build a public /mcus page from the same data `sim models` already ships: each chip's status (modeled, in progress, not modeled), the peripherals and buses modeled, the supported build systems and the sim version it arrived in. Give each released chip its own page (for example /mcus/stm32f401re) with a ten-line runnable example, the PR-review setup for that chip and its validation status. Add a "request a chip" button that asks for the part number and the use case. Link the list from the homepage hero ("Runs today on STM32F0/F4/H7, ESP32 and RP2040"), the Cloud Sim page, pricing and the PR-review docs, and date each new model so the weekly additions show as visible progress.

    It answers the qualifying question before signup, prevents wasted static reviews, and gives Simantic pages that match how engineers search ("STM32F401RE simulator", "simulate ESP32-C3 firmware"). Chip requests become a demand-ranked modeling roadmap. Renode leads with its public board list for the same reason.

    Ready to use

    ChecklistSupported-chip page launch list

    [ ] /mcus table generated from `sim models --show-unmodeled`
    [ ] Status per chip: modeled / in progress / not modeled
    [ ] Buses and peripherals modeled per chip
    [ ] Build systems: CubeMX make, make, CMake, pico-sdk, ESP-IDF, Zephyr
    [ ] One page per released chip with a runnable example
    [ ] "Request a chip" form (part number, use case)
    [ ] Hero line on the homepage linking to /mcus
    [ ] Links from /cloud-sim, /pricing and /docs/pr-review
    [ ] Static-review message links to /mcus
    [ ] Release date on each newly modeled chip
    [ ] Pages added to the sitemap

    Risks and how to handle them

    • The public list drifts from what the binary supports

      Generate it from the same source as `sim models` on every release.

    • A short list looks thin against Renode's 1200+ boards

      Lead with depth per chip (buses, timing, validation) and show requests and release dates.

    Evidence

    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • Your websiteCloud Sim page (/cloud-sim), the target of the homepage "Try Cloud Sim" button, checked twice on November 9, 2026: its whole body is "Introducing Cloud Sim" and "Cloud Sim lets you run your firmware from anywhere, with minimal setup." It names no chip and has no code sample or steps; the only next step is the Dashboard link.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
    • Your websiteGuide (/docs/guide): examples name STM32F401RE, STM32H753IITX, STM32G474RE, STM32F767ZI, NRF52840QIAA and ESP32-C3, C6 and S3, and point to `sim models` for the list; modeled buses include UART, I2C, SPI, CAN FD, BLE, 802.15.4, Ethernet, Wi-Fi, SD/MMC, USB, GPIO, timers and DMA; Python "peers" stand in for sensors and other parts; "`--timeout` is virtual seconds".Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Competitor researchRenode (Antmicro's platform page): an open-source emulation framework claiming 1200+ supported boards, 12+ CPU architectures and 800+ peripheral models, with public Zephyr and U-Boot test dashboards per board, integrations with GitHub Actions, GitLab and Jenkins, and paid services for peripheral modeling and CI setup.Source
  • 02

    "Try Cloud Sim" is the homepage's main product button, and it leads to a page whose whole body is one sentence and a Dashboard link. The docs already have everything a convincing page needs: a one-line curl install, `sim login`, `sim models` and a run command with `--expect` assertions for CI. The TypeScript SDK and MCP server are also live on npm.

    High impactLow effort First result in 2 weeks Founder/CTO

    First action

    Copy the docs quickstart and one real run's output onto /cloud-sim.

    How to measure it

    /cloud-sim to signup conversion

    Now
    –
    Target
    +50%
    By
    45 days

    Steps

    1. 1

      Pick the demo run

      One sample ELF on STM32F401RE with clear UART output.

      Founder/CTO 2 hours

    2. 2

      Write the quickstart

      Install, login, models, run and expect, with output.

      Founder/CTO 1 day

    3. 3

      Add interface tabs

      CLI, HTTP API, TypeScript and MCP versions of the same run.

      Founder/CTO 1 day

    4. 4

      Add proof points

      Determinism, virtual-time timeouts and buses modeled.

      Founder/CEO 2 hours

    5. 5

      One call to action

      Sign up for Community with the credit math beside it.

      Founder/CEO 1 hour

    6. 6

      Measure

      Page-to-signup and signup-to-first-run rates.

      Founder/CEO weekly

    First result
    2 weeks
    Cost
    Two days of founder time.
    Owner
    Founder/CTO

    Smallest test

    Paste the docs quickstart onto /cloud-sim today and compare signups week over week.

    Why it works

    Rebuild /cloud-sim as a quickstart: the install line, a run command against a sample ELF, the actual UART output and a pass/fail `--expect` line, then tabs for CLI, HTTP API, TypeScript and MCP showing the same run. Add the chip list (idea 1), what a run costs in credits (idea 3) and one signup button.

    Developer tools convert when visitors can see the exact command and its output before signing up. The page already gets the homepage's main click.

    Risks and how to handle them

    • Commands change between releases

      Pull snippets from the docs source so they update together.

    Evidence

    • Your websiteCloud Sim page (/cloud-sim), the target of the homepage "Try Cloud Sim" button, checked twice on November 9, 2026: its whole body is "Introducing Cloud Sim" and "Cloud Sim lets you run your firmware from anywhere, with minimal setup." It names no chip and has no code sample or steps; the only next step is the Dashboard link.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
    • npmTwo packages: @simantic/cloud-sim (35 weekly downloads) and @simantic/cloud-sim-mcp (29 weekly downloads), the TypeScript SDK and MCP server linked from the homepage.Source
  • 03

    The pricing page sells credits in dollars ($10 on Community, $200 on Developer) but never says what a simulation run or a PR review costs. Community shows no price at all, and there is no FAQ. The details are only in the PR-review docs: simulations and reviews share the allowance, and a review is refused with less than $1.50 left. An engineer can't tell whether $200 covers a team's CI or ten reviews. Wokwi, by contrast, sells "2000 monthly CI minutes".

    High impactLow effort First result in 2 weeks Founder/CEO

    First action

    Pull last month's median cost per review and per simulated minute from billing.

    How to measure it

    Pricing-to-Developer conversion

    Now
    –
    Target
    +30%
    By
    60 days

    Steps

    1. 1

      Measure unit costs

      Median and p90 cost per review and per simulated minute.

      Founder/CTO 1 day

    2. 2

      Write equivalents

      "About X reviews or Y minutes a month" per plan.

      Founder/CEO 2 hours

    3. 3

      Label Community

      Show "Free" and what needs no card.

      Founder/CEO 1 hour

    4. 4

      Add the FAQ

      Credits, overage budget, refusals under $1.50, annual savings.

      Founder/CEO half a day

    5. 5

      Add a calculator

      Reviews and CI runs per week in, recommended plan out.

      Founder/CTO 2 days

    6. 6

      Review quarterly

      Update equivalents as the engine gets faster.

      Founder/CEO quarterly

    First result
    2 weeks
    Cost
    $0; a few hours of founder time.
    Owner
    Founder/CEO

    Smallest test

    Add one line per plan, "about N PR reviews a month", this week.

    Why it works

    Under each plan, add typical usage: about N simulated minutes or M PR reviews a month, based on real median costs. Mark Community "Free". Add an FAQ: what a credit is, what a typical review and run cost, what happens when credits run out, how the overage budget works, and the difference between monthly and annual billing.

    Usage pricing in dollars without units makes engineers guess and stall. Concrete equivalents let them choose a plan, and they make the $1,000 annual Developer plan easier to justify to a manager.

    Risks and how to handle them

    • Costs vary widely by firmware

      Show a typical range and the p90.

    • Printing unit costs fixes expectations

      Call them estimates and date them.

    Evidence

    • Your websitePricing page: Community ("For individuals getting started", no price shown, "$10 in monthly credits", 3 concurrent simulation runs, GitHub integration, "Sign up"); Developer ("For professional engineers", $100 a month or $1,000 a year, shown as $83.33/mo, "$200 in monthly credits", 10 concurrent runs, early access to simulation models); Enterprise ("Contact Us": unlimited API usage, unlimited concurrent runs and PR reviews, private hosting, custom subdomain and SSO, custom SLA). There is no FAQ and no explanation of what a credit buys.Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Competitor researchWokwi (homepage and pricing): "World's most advanced ESP32 simulator", running in the browser for Arduino, ESP32, STM32 and Raspberry Pi Pico; Community is €0 with unlimited simulations; Wokwi Pro is €20 per seat a month with "2000 monthly CI minutes", VS Code and JetBrains plug-ins and team billing.Source
  • 04

    The homepage lists Cloud Sim access as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon". But /docs documents the `sim` CLI in full, with a one-line install, `sim update` and CI flags, and the public GitHub org has `pippy`, a Python SDK and pytest plugin (alpha, 0.4.x). Firmware test engineers who live in a terminal or pytest are told to come back later for tools that already exist.

    Medium impactLow effort First result in 1 week Founder/CTO

    First action

    Edit the two labels and links on the homepage.

    How to measure it

    Homepage-to-install clicks

    Now
    –
    Target
    +25%
    By
    30 days

    Steps

    1. 1

      Relabel the CLI

      Link to the /docs install section.

      Founder/CTO 30 minutes

    2. 2

      Relabel Python

      "Python (alpha)" linked to the pippy README.

      Founder/CTO 30 minutes

    3. 3

      Add a pytest snippet

      The timer-interrupt test from the README.

      Founder/CTO 1 hour

    4. 4

      Sweep the site

      Find other stale "coming soon" or outdated lines.

      Founder/CEO 1 hour

    5. 5

      Add to the release checklist

      Review homepage labels on each sim release.

      Founder/CTO 15 minutes

    6. 6

      Measure

      Clicks to the install and pippy links.

      Founder/CEO monthly

    First result
    1 week
    Cost
    $0
    Owner
    Founder/CTO

    Smallest test

    Change the CLI label today.

    Why it works

    Change "CLI (coming soon)" to "CLI" linked to the docs install, and "Python (coming soon)" to "Python (alpha)" linked to the pippy README, with a four-line pytest example. Check every other "coming soon" or stale claim on the site at each release.

    The CLI and Python are the interfaces embedded test engineers reach for first. Showing them costs nothing and removes a reason to leave.

    Risks and how to handle them

    • Alpha users hit breaking changes

      Label it alpha and pin versions, as the README already advises.

    Evidence

    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
  • 05

    Launch coverage promised CI that validates every commit, and the PR reviewer docs say plainly that it is "not CI: nothing runs on a push". The CI piece exists: `embedded-ci-action` on GitHub submits a built ELF for simulation and reports into the dashboard's Embedded CI tab. But it has no release (examples pin @main), no Marketplace badge, and the website has no CI page or link. Wokwi makes CI minutes a headline feature of its paid plan.

    High impactLow effort First result in 1 month Founder/CTO

    First action

    Tag v1.0.0 of embedded-ci-action and update the README to pin it.

    How to measure it

    Weekly active CI repositories

    Now
    –
    Target
    50
    By
    90 days

    Steps

    1. 1

      Tag a release

      v1 for the action and the setup step; pin it in the README.

      Founder/CTO half a day

    2. 2

      Publish to the Marketplace

      Branding, description and categories.

      Founder/CTO 2 hours

    3. 3

      Write /ci

      YAML, assertions, cost per run and a dashboard screenshot.

      Founder/CTO 1 day

    4. 4

      Link it

      Homepage, pricing and PR-review docs.

      Founder/CEO 1 hour

    5. 5

      Seed users

      Offer setup help to the first ten teams that sign up.

      Founder/CEO 2 weeks

    6. 6

      Track

      Active repos, runs per week and the upgrade rate.

      Founder/CEO weekly

    First result
    1 month
    Cost
    About two days of founder time.
    Owner
    Founder/CTO

    Smallest test

    Tag the release and add a CI section to /docs this week.

    Why it works

    Tag v1 of the action and its setup step, publish it to the GitHub Marketplace, and add a /ci page: the workflow YAML, what a CI run costs in credits, what fails a build (`--expect` and `--expect-absent`), and a screenshot of the Embedded CI tab. Link /ci from the homepage and pricing, and from the PR reviewer docs ("for every push, use CI").

    CI is a recurring use that consumes credits every day, which is what turns a $10 trial into a Developer or Team plan. The Marketplace is where firmware teams look for actions.

    Risks and how to handle them

    • CI runs exhaust Community credits quickly

      Show per-run cost and a warning before the allowance runs out.

    Evidence

    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Your websiteSitemap (14 URLs, checked November 9, 2026): the homepage, /cloud-sim, /pricing, /validation, /contact, /investors, /apply, /blog, /privacy, /terms and four docs pages (/docs, /docs/guide, /docs/pr-review, /docs/simantic-json). There are no supported-chip, CI, use-case, changelog or comparison pages, and the blog says only "We're working on it. Check back later."Source
    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Competitor researchWokwi (homepage and pricing): "World's most advanced ESP32 simulator", running in the browser for Arduino, ESP32, STM32 and Raspberry Pi Pico; Community is €0 with unlimited simulations; Wokwi Pro is €20 per seat a month with "2000 monthly CI minutes", VS Code and JetBrains plug-ins and team billing.Source
  • 06

    Simantic is recruiting a Lead Simulation Researcher at $200K-$300K with 2-5% equity, the hire that decides its electronics roadmap. The footer's "Careers" link goes to /apply, which shows readers that don't run scripts only "Open roles at Simantic" and "Loading…". The role text lives only on Y Combinator's job board, and /careers returns a 404.

    Medium impactLow effort First result in 2 weeks Founder/CEO

    First action

    Paste the Lead Simulation Researcher description into /apply as static text.

    How to measure it

    Qualified applicants

    Now
    –
    Target
    10
    By
    60 days

    Steps

    1. 1

      Render roles statically

      Build the role text into the page at deploy time.

      Founder/CTO half a day

    2. 2

      Add structured data

      JobPosting with location, salary range and remote status.

      Founder/CTO 2 hours

    3. 3

      Fix /careers

      Redirect it to /apply.

      Founder/CTO 15 minutes

    4. 4

      Write "why join"

      Engine, validation rigs, backers, Waterloo and San Francisco.

      Founder/CEO 2 hours

    5. 5

      Share it

      University of Waterloo and simulation communities, LinkedIn.

      Founder/CEO 1 week

    6. 6

      Track

      Applicants by source.

      Founder/CEO weekly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Add the role text to /apply today and check how it renders with scripts off.

    Why it works

    Render each open role into the /apply page as static HTML with the full description, add JobPosting structured data, redirect /careers to /apply, and link the Y Combinator listing. Add a short "why join" section: the from-scratch Rust engine, the hardware validation rigs and the backers.

    Senior simulation researchers are rare and research a company before applying. A page that search engines and job aggregators can read widens the funnel for the most important hire.

    Risks and how to handle them

    • Roles go stale on the page

      Generate the page from the same source as the job board.

    Evidence

    • Your websiteCareers page (/apply, the footer's "Careers" link), read November 9, 2026: a reader that does not run scripts sees only "Join us", "Open roles at Simantic" and "Loading…"; no role title or description is in the page. /careers returns a 404.Source
    • HiringLead Simulation Researcher (Electronics / Firmware), Waterloo or San Francisco, $200K-$300K and 2.00%-5.00% equity, 6+ years: to "lead Simantic's efforts to develop an electronics simulation platform based on SPICE and Xyce for agentic AI use", prototyping hardware to check that the simulation physics matches real-world behavior. The listing describes a team of 4.Source
  • 07

    The launch post asks to hear from teams in automotive, robotics and aerospace, and Enterprise is contact-only. But the contact form asks only for name, job title, company, work email and "How can we help?". It has no booking link, no email address and no question about the chip, industry or CI system, so every enterprise lead needs a round of email before anyone can judge fit.

    Medium impactLow effort First result in 2 weeks Founder/CEO

    First action

    Add the MCU dropdown and a calendar link to the form.

    How to measure it

    Form-to-call rate

    Now
    –
    Target
    60%
    By
    45 days

    Steps

    1. 1

      Add fields

      MCU, industry, CI system.

      Founder/CTO 2 hours

    2. 2

      Add booking

      Calendar link on the thank-you page.

      Founder/CEO 1 hour

    3. 3

      Auto-reply

      Send the relevant chip page or the nearest supported part.

      Founder/CTO half a day

    4. 4

      Route requests

      Unsupported chips go to the modeling backlog.

      Founder/CTO 1 hour

    5. 5

      Review weekly

      Which industries and chips ask most.

      Founder/CEO weekly

    6. 6

      Measure

      Form-to-call and call-to-pilot rates.

      Founder/CEO monthly

    First result
    2 weeks
    Cost
    $0-$15 a month for a scheduling tool.
    Owner
    Founder/CEO

    Smallest test

    Add the calendar link to the thank-you page today.

    Why it works

    Add three short fields: target MCU (a dropdown from the chip list, plus "other"), industry, and CI system. Show a calendar link on the thank-you page. Route "other" chips to the modeling backlog and send an instant reply with the matching chip page or the nearest supported part.

    Qualifying before the call shortens the cycle, and a booking link captures intent while the visitor is still on the page.

    Risks and how to handle them

    • More fields lower submissions

      Keep them optional dropdowns.

    Builds on

    Evidence

    • Your websiteContact page ("Modernize your team's firmware workflows"): a form for first and last name, job title, company, work email and "How can we help?"; no booking link, no email address and no question about the target chip, industry or CI system.Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
    • Your websitePricing page: Community ("For individuals getting started", no price shown, "$10 in monthly credits", 3 concurrent simulation runs, GitHub integration, "Sign up"); Developer ("For professional engineers", $100 a month or $1,000 a year, shown as $83.33/mo, "$200 in monthly credits", 10 concurrent runs, early access to simulation models); Enterprise ("Contact Us": unlimited API usage, unlimited concurrent runs and PR reviews, private hosting, custom subdomain and SSO, custom SLA). There is no FAQ and no explanation of what a credit buys.Source

This quarter 11

The core plays for the next three months.

  • 08

    The validation page promises to "Show the evidence. Name the boundaries." and to build "a validation standard that firmware simulation has not seen before", yet it names no chips and shows no numbers. The company says it runs hardware validation rigs around the clock. Buyers in automotive, robotics and aerospace, the teams the launch asks for, will not trust a simulator without evidence.

    High impactMedium effort First result in 6 weeks Founder/CTO

    First action

    Pick the best-validated chip and write its first report from existing rig data.

    How to measure it

    Enterprise inquiries

    Now
    –
    Target
    10 a month
    By
    1 quarter

    Steps

    1. 1

      Define the report

      Suites, metrics, version and known gaps.

      Founder/CTO 1 week

    2. 2

      Export rig results

      Pass and mismatch counts per chip and suite.

      Founder/CTO 1 week

    3. 3

      Write the first report

      The best-validated chip first.

      Founder/CTO 1 week

    4. 4

      Publish and badge

      On /validation and the chip page.

      Founder/CTO 2 days

    5. 5

      Automate

      Regenerate results on each sim release.

      Founder/CTO 2 weeks

    6. 6

      Use in sales

      Attach the report to every enterprise conversation.

      Founder/CEO ongoing

    First result
    6 weeks
    Cost
    Engineering time; rigs already run.
    Owner
    Founder/CTO

    Smallest test

    Publish one chip's results table on /validation.

    Why it works

    For each released chip, publish a dated validation report: the test suite (instructions, peripherals, timing, multi-device scenarios), the pass rate against real silicon, the known simplifications and any open mismatches, tied to the sim version. Put a status badge on each chip page from idea 1 and offer a downloadable report for enterprise security and quality reviews.

    Accuracy is the deciding question for simulation. Publishing it, including the gaps, turns the validation principle into proof that competitors don't show per chip.

    Risks and how to handle them

    • Published gaps are used against Simantic

      Pair each gap with its fix date; the honesty is the differentiator.

    Builds on

    Evidence

    • Your websiteValidation page: principles such as "Let real hardware be the reference", "Look beyond a successful boot", "Make failures part of the test" and "Show the evidence. Name the boundaries.", and "We are building toward a validation standard that firmware simulation has not seen before." It names no chips and shows no numbers or results; the call to action is "Talk to the team".Source
    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
    • Competitor researchRenode (Antmicro's platform page): an open-source emulation framework claiming 1200+ supported boards, 12+ CPU architectures and 800+ peripheral models, with public Zephyr and U-Boot test dashboards per board, integrations with GitHub Actions, GitLab and Jenkins, and paid services for peripheral modeling and CI setup.Source
  • 09

    The blog says only "We're working on it. Check back later.", and the sitemap has no changelog. Meanwhile the founders say new models arrive weekly, and the docs record milestones such as the STM32H743/H753 parts "modeled since sim 0.8". The most persuasive content Simantic has, steady product progress, is invisible.

    Medium impactLow effort First result in 1 month Founder/CEO

    First action

    Write the first entry covering the launch chips and sim 0.8's H7 parts.

    How to measure it

    Email and RSS subscribers

    Now
    –
    Target
    500
    By
    1 quarter

    Steps

    1. 1

      Create /changelog

      Astro content collection with RSS.

      Founder/CTO half a day

    2. 2

      Backfill

      Launch chips and recent releases.

      Founder/CEO 1 day

    3. 3

      Set a cadence

      One entry every Friday.

      Founder/CEO weekly

    4. 4

      Add email signup

      On the changelog and docs.

      Founder/CTO 2 hours

    5. 5

      Write deep dives

      One engineering post a month.

      Founder/CTO monthly

    6. 6

      Distribute

      LinkedIn and embedded communities.

      Founder/CEO weekly

    First result
    1 month
    Cost
    An hour a week.
    Owner
    Founder/CEO

    Smallest test

    Replace the blog placeholder with one launch-week changelog entry.

    Why it works

    Publish a short weekly post or /changelog entry: new chips and peripherals modeled, validation changes, CLI and SDK releases, and one example run. Offer an RSS feed and an email digest, and repost each entry to LinkedIn. Once a month, write one deeper engineering post (how a peripheral was modeled, or a timing bug found against silicon).

    Weekly releases show a young company is moving, give returning visitors a reason to come back, and build search pages for each newly supported part.

    Risks and how to handle them

    • Cadence slips

      Generate a draft from release notes automatically.

    Evidence

    • Your websiteSitemap (14 URLs, checked November 9, 2026): the homepage, /cloud-sim, /pricing, /validation, /contact, /investors, /apply, /blog, /privacy, /terms and four docs pages (/docs, /docs/guide, /docs/pr-review, /docs/simantic-json). There are no supported-chip, CI, use-case, changelog or comparison pages, and the blog says only "We're working on it. Check back later."Source
    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
  • 10

    The homepage's proof is a sample review of a toy STM32F072RB blinky. Simantic's GitHub org already forks Zephyr and open-source flight-controller, e-reader, solar-car and student racing firmware, and reviews can be shared by link. A real bug caught in a project engineers recognise is far stronger proof than a demo.

    High impactMedium effort First result in 1 month Founder/CTO

    First action

    Pick three active open-source projects on supported chips and run the bot on their last ten merged PRs.

    How to measure it

    Bugs confirmed by maintainers

    Now
    –
    Target
    10
    By
    1 quarter

    Steps

    1. 1

      Choose projects

      Active repos on STM32, ESP32 or RP2040.

      Founder/CTO 1 day

    2. 2

      Run reviews

      Recent merged and open PRs.

      Founder/CTO 1 week

    3. 3

      Verify findings

      Reproduce each one before contacting anyone.

      Founder/CTO 1 week

    4. 4

      Report upstream

      Through the project's issue process, respectfully.

      Founder/CTO ongoing

    5. 5

      Publish with consent

      A "found in the wild" page with shared review links.

      Founder/CEO 2 weeks

    6. 6

      Offer free reviews

      To maintainers of those projects.

      Founder/CEO ongoing

    7. 7

      Measure

      Confirmed bugs, referral signups and stars.

      Founder/CEO monthly

    First result
    1 month
    Cost
    Credits for reviews and founder time.
    Owner
    Founder/CTO

    Smallest test

    Review ten merged PRs in one project and count genuine findings.

    Why it works

    Run Simantic Bot on recent pull requests in public firmware projects that target supported chips. When it finds a genuine issue, confirm it, report it upstream through the project's normal process, and, with the maintainers' agreement, publish a "found in the wild" page: the bug, the simulator evidence and the fix. Offer free reviews to the maintainers of those projects.

    It shows the product working on code people know, earns goodwill and links from open-source communities, and produces a steady stream of technical stories for the changelog and social posts.

    Risks and how to handle them

    • False positives annoy maintainers

      Only report findings that are reproduced by hand.

    • Projects object to publicity

      Publish only with consent; otherwise keep it private.

    Evidence

    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
  • 11

    Simantic grew out of the University of Waterloo and its Velocity programme, and its GitHub org already forks student solar-car and racing-team firmware. These teams fight for boards and bench time, and their members graduate into the automotive, robotics and aerospace companies the launch targets.

    Medium impactLow effort First result in 1 month Founder/CEO

    First action

    Offer the programme to three Waterloo student design teams.

    How to measure it

    Active student teams

    Now
    –
    Target
    15
    By
    2 quarters

    Steps

    1. 1

      Define the offer

      Credits, CI, support and the case-study ask.

      Founder/CEO 1 day

    2. 2

      Start locally

      Three Waterloo design teams.

      Founder/CEO 2 weeks

    3. 3

      Onboard

      simantic.json and CI on their main repos.

      Founder/CTO 1 week per team

    4. 4

      Collect stories

      One case study per team.

      Founder/CEO 1 quarter

    5. 5

      Expand

      Other universities and competitions.

      Founder/CEO 1 quarter

    6. 6

      Track

      Active teams, runs and later conversion of alumni.

      Founder/CEO monthly

    First result
    1 month
    Cost
    Credits worth a few hundred dollars a month.
    Owner
    Founder/CEO

    Smallest test

    Onboard one team's firmware repo this month.

    Why it works

    Offer competition teams boosted Community credits, a shared org install of Simantic Bot and CI, and a channel for support, in exchange for a short case study and feedback on missing peripherals. Start with Waterloo design teams, then student engineering competitions across North America.

    It builds a cohort of engineers who learn on Simantic and carry it into industry, while producing real-world case studies and modeling requests from demanding firmware.

    Risks and how to handle them

    • Teams use chips not yet modeled

      Pick teams on supported chips first; log the rest as requests.

    Builds on

    Evidence

    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
    • News coverageUniversity of Waterloo news (July 31, 2026): the founders met the problem in co-op roles where engineers often waited weeks or months for physical devices; the company started in September 2025 and was accepted into Velocity's startup program and Y Combinator's Fall 2026 batch; it has raised about $600,000 USD; the founders met customers, design partners and investors in San Francisco.Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
  • 12

    pippy already lets engineers drive simulated firmware from pytest, with fixtures, UART expectations and virtual time. But it is alpha (0.4.x), "not distributed on PyPI", may change "without a deprecation period" and advises against production pipelines. The homepage still calls Python "coming soon". Test engineers standardise on pytest, so this is the shortest path into their existing suites.

    Medium impactMedium effort First result in 6 weeks Founder/CTO

    First action

    List the API surface to freeze for 1.0.

    How to measure it

    PyPI downloads per week

    Now
    –
    Target
    500
    By
    1 quarter

    Steps

    1. 1

      Freeze the API

      Fixture, expect, read and run_firmware.

      Founder/CTO 1 week

    2. 2

      Package

      Wheels for linux-x64 and osx-arm64.

      Founder/CTO 1 week

    3. 3

      Publish 1.0

      PyPI with a changelog.

      Founder/CTO 1 day

    4. 4

      Write recipes

      Three pytest examples in the docs.

      Founder/CTO 2 days

    5. 5

      Link it

      Homepage, /cloud-sim and the CI page.

      Founder/CEO 1 hour

    6. 6

      Measure

      Installs and first runs per week.

      Founder/CEO weekly

    First result
    6 weeks
    Cost
    Two to three weeks of engineering.
    Owner
    Founder/CTO

    Smallest test

    Publish a pre-release to PyPI and invite five users.

    Why it works

    Freeze a 1.0 API for the pytest fixture and run_firmware, publish to PyPI with semantic versioning, add a docs page with three recipes (UART expectation, register read, faulty-peer test) and link it from the homepage and /cloud-sim.

    `pip install` plus a pytest fixture is a familiar, low-friction start, and a stable API is what lets teams put it in CI.

    Risks and how to handle them

    • Freezing the API too early

      Freeze only the small core; keep the rest experimental.

    Builds on

    Evidence

    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
  • 13

    Launch coverage already says Simantic built its engine from scratch rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy. Engineers evaluating simulation search for exactly those names. Renode claims 1200+ boards and Wokwi offers free browser simulation with CI in Pro, but simantic.com has no comparison pages.

    Medium impactMedium effort First result in 6 weeks Founder/CEO

    First action

    Draft the Renode comparison from the engineering reasons the founders already give.

    How to measure it

    Comparison-page signups

    Now
    –
    Target
    40 a month
    By
    1 quarter

    Steps

    1. 1

      Gather facts

      Each tool's current docs, pricing and board lists.

      Founder/CEO 2 days

    2. 2

      Benchmark

      Same firmware, run time and timing results.

      Founder/CTO 1 week

    3. 3

      Write pages

      Fair tables, "choose them if" sections and migration notes.

      Founder/CEO 1 week

    4. 4

      Review claims

      Check every competitor statement against its own pages.

      Founder/CTO 2 days

    5. 5

      Publish

      Linked from the chip pages and docs.

      Founder/CTO 1 day

    6. 6

      Refresh

      Every quarter.

      Founder/CEO quarterly

    First result
    6 weeks
    Cost
    Founder time only.
    Owner
    Founder/CEO

    Smallest test

    Publish the Renode page and watch search impressions for a month.

    Why it works

    Write one honest page per alternative: what each is best at (Renode's board breadth, QEMU's ubiquity, Wokwi's browser and classroom reach), where Simantic differs (determinism, virtual time, agent and MCP access, simulation-checked PR review, published validation) and how to move a test suite across. Date each page and cite sources.

    Comparison searches come from engineers already choosing a tool. Honest pages earn trust and help AI assistants describe Simantic correctly.

    Risks and how to handle them

    • Seen as attacking open-source projects

      Credit them generously and state where they are the better choice.

    Builds on

    Evidence

    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Competitor researchWokwi (homepage and pricing): "World's most advanced ESP32 simulator", running in the browser for Arduino, ESP32, STM32 and Raspberry Pi Pico; Community is €0 with unlimited simulations; Wokwi Pro is €20 per seat a month with "2000 monthly CI minutes", VS Code and JetBrains plug-ins and team billing.Source
    • Competitor researchRenode (Antmicro's platform page): an open-source emulation framework claiming 1200+ supported boards, 12+ CPU architectures and 800+ peripheral models, with public Zephyr and U-Boot test dashboards per board, integrations with GitHub Actions, GitLab and Jenkins, and paid services for peripheral modeling and CI setup.Source
    • Your websiteSitemap (14 URLs, checked November 9, 2026): the homepage, /cloud-sim, /pricing, /validation, /contact, /investors, /apply, /blog, /privacy, /terms and four docs pages (/docs, /docs/guide, /docs/pr-review, /docs/simantic-json). There are no supported-chip, CI, use-case, changelog or comparison pages, and the blog says only "We're working on it. Check back later."Source
  • 14

    Simantic positions itself as "the world's first agent-first firmware simulator" and ships an MCP server, but @simantic/cloud-sim-mcp has 29 weekly downloads and @simantic/cloud-sim 35. Engineers wiring coding agents into firmware work find tools through MCP registries, editor docs and example repos, and Simantic isn't visible there yet.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Rewrite the cloud-sim-mcp README with a tools table and a config snippet.

    How to measure it

    cloud-sim-mcp weekly downloads

    Now
    –
    Target
    300
    By
    1 quarter

    Steps

    1. 1

      Improve the README

      Tools, config and a demo GIF.

      Founder/CTO 1 day

    2. 2

      Submit to registries

      Main MCP registries and directories.

      Founder/CTO 1 day

    3. 3

      Write /docs/agents

      Agent loop: write, run, assert, fix.

      Founder/CTO 2 days

    4. 4

      Publish a template

      Repo with simantic.json, CI and MCP config.

      Founder/CTO 2 days

    5. 5

      Record a demo

      Two-minute video of an agent fixing a real bug.

      Founder/CEO 1 day

    6. 6

      Track

      Downloads, signups with an MCP source.

      Founder/CEO weekly

    First result
    1 month
    Cost
    About a week of founder time.
    Owner
    Founder/CTO

    Smallest test

    Submit to one registry and compare downloads after two weeks.

    Why it works

    Write a clear README for the MCP package (the tools it exposes, a config snippet, a 30-second demo GIF), submit it to the main MCP registries and directories, and add a /docs/agents page showing an agent writing, running and fixing firmware against Cloud Sim. Publish a template repo with simantic.json, CI and the MCP config ready to use.

    Agent builders adopt tools that are one config line away. Being listed where they search is cheap distribution for the exact audience the launch names.

    Risks and how to handle them

    • Agent users burn credits quickly

      Default the MCP to short virtual timeouts and show cost per call.

    Evidence

    • npmTwo packages: @simantic/cloud-sim (35 weekly downloads) and @simantic/cloud-sim-mcp (29 weekly downloads), the TypeScript SDK and MCP server linked from the homepage.Source
    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
  • 15

    The docs say "Usage reporting is required in this build and has no opt-out": each command reports its options, duration and, for runs, the ELF's size and SHA-256. PR reviews send the diff, build output and simulator output to a model provider. Enterprise promises private hosting, yet there is no security or data page, and a hardware-first rival, Embedder, advertises SOC 2 Type II, ISO 27001 and air-gapped deployment. For the automotive and aerospace teams Simantic wants, a mandatory-telemetry line can stall a security review.

    High impactMedium effort First result in 1 quarter Founder/CEO

    First action

    Write the "what leaves your machine" table from the existing docs.

    How to measure it

    Enterprise pilots past security review

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Map data flows

      CLI, API, CI and PR review, item by item.

      Founder/CTO 2 days

    2. 2

      Write /security

      Flows, retention, subprocessors and controls.

      Founder/CEO 3 days

    3. 3

      Build the opt-out

      Telemetry off for Enterprise and private hosting.

      Founder/CTO 2 weeks

    4. 4

      Set the roadmap

      SOC 2 readiness timeline.

      Founder/CEO 1 week

    5. 5

      Make a summary

      One-page PDF for procurement.

      Founder/CEO 1 day

    6. 6

      Link it

      Pricing Enterprise column and the contact form.

      Founder/CTO 1 hour

    First result
    1 quarter
    Cost
    Founder time; compliance tooling later.
    Owner
    Founder/CEO

    Smallest test

    Publish the data-flow table and send it to one enterprise prospect.

    Why it works

    Publish /security: exactly what leaves the machine for each product, retention periods, subprocessors, where reviews run, the firmware-storage switch, and the compliance roadmap. Offer a telemetry opt-out (or offline licence) for Enterprise and private-hosting builds, and a one-page security summary for procurement.

    Regulated buyers move faster when the answers to their first questionnaire are already public. Turning "no opt-out" into an Enterprise feature gives a concrete reason to upgrade.

    Risks and how to handle them

    • Less usage data from big accounts

      Keep aggregate, opt-in reporting for Enterprise.

    Evidence

    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Your websitePricing page: Community ("For individuals getting started", no price shown, "$10 in monthly credits", 3 concurrent simulation runs, GitHub integration, "Sign up"); Developer ("For professional engineers", $100 a month or $1,000 a year, shown as $83.33/mo, "$200 in monthly credits", 10 concurrent runs, early access to simulation models); Enterprise ("Contact Us": unlimited API usage, unlimited concurrent runs and PR reviews, private hosting, custom subdomain and SSO, custom SLA). There is no FAQ and no explanation of what a credit buys.Source
    • Competitor researchEmbedder (homepage): a Y Combinator-backed "autonomous firmware engineer" claiming 10,000+ developers; it tests on real silicon, probes and instruments rather than simulation; it states SOC 2 Type II and ISO 27001, with on-prem and air-gapped deployment options.Source
  • 16

    Every path into Simantic starts with your own compiled ELF: Cloud Sim needs "an MCU model and a compiled ELF", and the public examples repo of good and faulty bus peers says "No firmware is included". A newcomer without an ARM or ESP-IDF toolchain set up can't see a single run, which is the moment the product sells itself.

    High impactLow effort First result in 1 month Founder/CTO

    First action

    Build and host three sample ELFs for STM32F401RE, RP2040 and ESP32-C3.

    How to measure it

    Day-1 first-run rate

    Now
    –
    Target
    60%
    By
    60 days

    Steps

    1. 1

      Build samples

      Four small firmwares per released chip.

      Founder/CTO 1 week

    2. 2

      Add the flag

      `sim run --sample` fetches and runs one.

      Founder/CTO 2 days

    3. 3

      Add a dashboard button

      "Run a sample" with live output.

      Founder/CTO 3 days

    4. 4

      Pair with faulty peers

      Show a bug being caught.

      Founder/CTO 2 days

    5. 5

      Update onboarding

      First email and quickstart point to samples.

      Founder/CEO 1 day

    6. 6

      Measure

      Time to first run and day-7 retention.

      Founder/CEO weekly

    First result
    1 month
    Cost
    About two weeks of engineering.
    Owner
    Founder/CTO

    Smallest test

    Add one downloadable sample ELF to the quickstart and compare first-run rates.

    Why it works

    Host sample ELFs per supported chip (blinky, UART hello, an I2C sensor read, a CAN node) and add `sim run --sample <name>` plus a "Run a sample" button in the dashboard. Pair each sample with its faulty-peer scenario from the examples repo so users see a firmware bug caught within minutes.

    Time to first successful run is the activation metric for developer tools. Removing the toolchain step turns curious signups into users who have seen it work.

    Risks and how to handle them

    • Samples feel like toys

      Include one realistic sensor and CAN example per chip.

    Builds on

    Evidence

    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
    • Your websiteCloud Sim page (/cloud-sim), the target of the homepage "Try Cloud Sim" button, checked twice on November 9, 2026: its whole body is "Introducing Cloud Sim" and "Cloud Sim lets you run your firmware from anywhere, with minimal setup." It names no chip and has no code sample or steps; the only next step is the Dashboard link.Source
  • 17

    Without a simantic.json, the reviewer detects the build and chip and "ends its review by suggesting you commit a simantic.json". A review turns static, and "can never be an approval", when the chip, toolchain or build can't be resolved. Asking the user to write the file by hand leaves the most common failure in place.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Count how many reviews last month were static because of build or chip detection.

    How to measure it

    Simulated (non-static) review rate

    Now
    –
    Target
    85%
    By
    60 days

    Steps

    1. 1

      Measure static reviews

      By cause: chip, toolchain, build, never ran.

      Founder/CTO 1 day

    2. 2

      Generate the file

      From the detected build and chip.

      Founder/CTO 3 days

    3. 3

      Open the PR

      On the add-config command.

      Founder/CTO 3 days

    4. 4

      Explain failures

      Name the missing field in the review.

      Founder/CTO 2 days

    5. 5

      Announce

      Changelog and docs.

      Founder/CEO 1 hour

    6. 6

      Track

      Static rate and config PR merges.

      Founder/CTO weekly

    First result
    1 month
    Cost
    About a week of engineering.
    Owner
    Founder/CTO

    Smallest test

    Paste the inferred simantic.json into the review body as a suggestion block.

    Why it works

    When the bot infers a working configuration, offer "@simantic-bot add-config", which opens a pull request containing the detected simantic.json (mcu, build system, ELF path, a smoke `expect`). When detection fails, the review explains which field is missing and links to the chip page.

    One click to a correct config makes every later review a full simulated review, which is when the product shows its value and uses credits.

    Risks and how to handle them

    • Wrong inferred config

      Open it as a PR for the user to review, never a direct commit.

    Evidence

    • Your websitesimantic.json docs: only `mcu` is required; without the file "the reviewer detects the build and MCU itself and ends its review by suggesting you commit a simantic.json"; supported build systems are CubeMX make, make, CMake, pico-sdk, ESP-IDF and Zephyr; a review is static and "can never be an approval" when the MCU is unsupported, the toolchain is missing, the build fails or the firmware never runs.Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
  • 18

    Developer is priced for one "professional engineer" at $100 a month, and the next step is a contact-only Enterprise plan. The PR-review docs already describe enterprise groups billed as a group, which is the mechanism a team plan needs. Wokwi sells Pro per seat with team billing and invoices. A five-person firmware team today has to choose between five separate accounts and a sales call.

    High impactMedium effort First result in 1 quarter Founder/CEO

    First action

    Find accounts that share an email domain and show them a Team offer.

    How to measure it

    Monthly recurring revenue from Team

    Now
    –
    Target
    20% of paid revenue
    By
    2 quarters

    Steps

    1. 1

      Size demand

      Accounts sharing a company domain.

      Founder/CEO 1 day

    2. 2

      Set pricing

      Per seat with pooled credits.

      Founder/CEO 1 week

    3. 3

      Reuse group billing

      Extend enterprise groups to self-serve.

      Founder/CTO 2 weeks

    4. 4

      Add admin views

      Seats, usage and invoices.

      Founder/CTO 1 week

    5. 5

      Launch

      Pricing page and an email to domain clusters.

      Founder/CEO 1 day

    6. 6

      Review

      Upgrades, seat expansion and churn.

      Founder/CEO monthly

    First result
    1 quarter
    Cost
    About a month of engineering.
    Owner
    Founder/CEO

    Smallest test

    Offer a manual team invoice to two companies with several accounts.

    Why it works

    Launch a self-serve Team plan: per-seat pricing, a pooled credit balance, one GitHub organization install for Simantic Bot and CI, higher concurrency, invoices and an admin view of usage. Keep Enterprise for private hosting, SSO and SLAs.

    Firmware work is done in teams, and a pooled, self-serve plan captures them before they need procurement, raising revenue per account.

    Risks and how to handle them

    • Cannibalizes Enterprise

      Keep private hosting, SSO and SLA Enterprise-only.

    Builds on

    Evidence

    • Your websitePricing page: Community ("For individuals getting started", no price shown, "$10 in monthly credits", 3 concurrent simulation runs, GitHub integration, "Sign up"); Developer ("For professional engineers", $100 a month or $1,000 a year, shown as $83.33/mo, "$200 in monthly credits", 10 concurrent runs, early access to simulation models); Enterprise ("Contact Us": unlimited API usage, unlimited concurrent runs and PR reviews, private hosting, custom subdomain and SSO, custom SLA). There is no FAQ and no explanation of what a credit buys.Source
    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Competitor researchWokwi (homepage and pricing): "World's most advanced ESP32 simulator", running in the browser for Arduino, ESP32, STM32 and Raspberry Pi Pico; Community is €0 with unlimited simulations; Wokwi Pro is €20 per seat a month with "2000 monthly CI minutes", VS Code and JetBrains plug-ins and team billing.Source

Big bets 3

Larger plays that can change the trajectory.

  • 19

    Simantic is hiring a Lead Simulation Researcher to build "an electronics simulation platform based on SPICE and Xyce for agentic AI use" and to check that simulation physics matches real hardware. Today's docs model the MCU, its buses and scripted peers. The homepage hero already shows the board-level questions engineers care about: I2C rise times and RC delays. Board-level faults (weak pull-ups, slow edges, brown-outs) are the ones firmware-only simulators miss.

    High impactHigh effort First result in 2 quarters Lead Simulation Researcher

    First action

    Write the brief for the first scenario and use it in interviews for the researcher role.

    How to measure it

    Add-on accounts

    Now
    –
    Target
    10
    By
    3 quarters

    Steps

    1. 1

      Hire the lead

      Use the first scenario as the interview problem.

      Founder/CEO 1 quarter

    2. 2

      Choose the scenario

      With three design partners.

      Lead Simulation Researcher 2 weeks

    3. 3

      Couple the models

      SPICE nets driven by MCU pins in virtual time.

      Lead Simulation Researcher 1 quarter

    4. 4

      Validate

      Against a real board on the rigs.

      Lead Simulation Researcher 1 month

    5. 5

      Publish results

      On /validation with known limits.

      Founder/CTO 2 weeks

    6. 6

      Price the add-on

      Per run or per seat.

      Founder/CEO 2 weeks

    7. 7

      Launch to partners

      Then general availability.

      Founder/CEO 1 quarter

    First result
    2 quarters
    Cost
    The researcher's salary plus rig hardware.
    Owner
    Lead Simulation Researcher

    Smallest test

    Prototype an I2C rise-time check for one board and show it to three design partners.

    Why it works

    Pick one board-level scenario with clear value, such as I2C pull-up and bus-capacitance checks, or reset and brown-out behaviour, and couple the SPICE model to the MCU simulation for that scenario only. Validate it against a real board, publish the results, and sell it as an add-on to Developer, Team and Enterprise.

    Joint firmware and electronics simulation is a gap neither firmware-only simulators nor hardware-first agents fill, and it fits the regulated buyers who pay most.

    Risks and how to handle them

    • Research takes longer than planned

      Ship one narrow, validated scenario before broadening.

    • Slow simulation

      Simulate only the nets that the scenario needs.

    Builds on

    Evidence

    • HiringLead Simulation Researcher (Electronics / Firmware), Waterloo or San Francisco, $200K-$300K and 2.00%-5.00% equity, 6+ years: to "lead Simantic's efforts to develop an electronics simulation platform based on SPICE and Xyce for agentic AI use", prototyping hardware to check that the simulation physics matches real-world behavior. The listing describes a team of 4.Source
    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • Your websiteGuide (/docs/guide): examples name STM32F401RE, STM32H753IITX, STM32G474RE, STM32F767ZI, NRF52840QIAA and ESP32-C3, C6 and S3, and point to `sim models` for the list; modeled buses include UART, I2C, SPI, CAN FD, BLE, 802.15.4, Ethernet, Wi-Fi, SD/MMC, USB, GPIO, timers and DMA; Python "peers" stand in for sensors and other parts; "`--timeout` is virtual seconds".Source
    • Your websiteValidation page: principles such as "Let real hardware be the reference", "Look beyond a successful boot", "Make failures part of the test" and "Show the evidence. Name the boundaries.", and "We are building toward a validation standard that firmware simulation has not seen before." It names no chips and shows no numbers or results; the call to action is "Talk to the team".Source
  • 20

    Simantic's founding insight is that engineers wait weeks or months for physical devices, and it says new chip models arrive weekly. Chip makers and distributors face the same wait at a new part's launch: engineers can't evaluate it until dev kits ship. A validated simulator model is a way to try a part in minutes.

    High impactHigh effort First result in 2 quarters Founder/CEO

    First action

    Build a polished demo page for one released chip to show in partner conversations.

    How to measure it

    Signed partnerships

    Now
    –
    Target
    2
    By
    3 quarters

    Steps

    1. 1

      Build the demo

      A "try it now" page for a released chip.

      Founder/CTO 2 weeks

    2. 2

      Package validation

      The chip's silicon-versus-simulation report.

      Founder/CTO 2 weeks

    3. 3

      List targets

      Vendor developer-relations teams and distributors for modeled families.

      Founder/CEO 1 week

    4. 4

      Pitch

      Use investor and YC introductions.

      Founder/CEO 1 quarter

    5. 5

      Pilot one launch

      A co-branded page for one part.

      Founder/CTO 1 month

    6. 6

      Measure

      Runs, signups and partner leads.

      Founder/CEO monthly

    First result
    2 quarters
    Cost
    Founder time and modeling effort per part.
    Owner
    Founder/CEO

    Smallest test

    Show the demo page to two vendor developer-relations contacts and gauge interest.

    Why it works

    Pitch chip makers' developer-relations teams and large distributors a co-branded "try it now" page for a new or popular part: a hosted sample, a Cloud Sim run in the browser and a path to Simantic for testing. Charge for priority modeling and validation of their parts, or run it as a lead-sharing partnership.

    It puts Simantic in front of engineers at the moment they evaluate a chip, under a trusted vendor's brand, and funds the modeling roadmap.

    Risks and how to handle them

    • Vendors demand accuracy proof

      Lead with the published validation report (idea 8).

    • Long partner cycles

      Start with a distributor or a smaller vendor.

    Builds on

    Evidence

    • News coverageUniversity of Waterloo news (July 31, 2026): the founders met the problem in co-op roles where engineers often waited weeks or months for physical devices; the company started in September 2025 and was accepted into Velocity's startup program and Y Combinator's Fall 2026 batch; it has raised about $600,000 USD; the founders met customers, design partners and investors in San Francisco.Source
    • News coverageAI Weekly (October 30, 2026): the engine was written from scratch in Rust rather than wrapping Renode or QEMU, citing development speed, performance and timing accuracy; the platform offers virtual dev boards by API, simulation-checked AI code review and CI that validates every commit; launch chips are STM32F0/F4/H7, ESP32-S3/C3 and RP2040, with new models arriving weekly; the company cites hardware validation rigs running around the clock; target uses include flight controllers, EVs, robotics and satellites.Source
    • Your websiteDocs (/docs): the `sim` CLI installs with one curl command on macOS (Apple Silicon) or Linux x64, and on Windows only through WSL2; it needs Python 3.12+ with its shared library, and every run needs a Simantic account. The sample `sim models` output lists RP2040, STM32F072RB, STM32F401RE and "..."; the STM32H743 and STM32H753 parts are "modeled since sim 0.8" while other H7 parts are "not-modeled". "Usage reporting is required in this build and has no opt-out": each command reports its options, duration, OS and, for runs, the ELF size and SHA-256 (firmware and UART output are not sent).Source
  • 21

    Pyrite, Simantic's simulation model, is "available to select partners" and can run locally in CI/CD or agent harnesses. The launch asks for automotive, robotics and aerospace teams, and Enterprise already lists private hosting and SLAs. But there is no public offer for Pyrite, and the validation page shows no results such buyers could take to a quality review.

    High impactHigh effort First result in 2 quarters Founder/CEO

    First action

    Write a one-page Pyrite licence offer and send it to current design partners.

    How to measure it

    Signed Pyrite licences

    Now
    –
    Target
    3
    By
    3 quarters

    Steps

    1. 1

      Define the offer

      Licence, validation pack, onboarding and SLA.

      Founder/CEO 1 week

    2. 2

      Add offline mode

      Licence file instead of per-run checks.

      Founder/CTO 1 month

    3. 3

      Build validation packs

      Per chip, from idea 8's reports.

      Founder/CTO 1 month

    4. 4

      Sign design partners

      Three, discounted for case studies.

      Founder/CEO 1 quarter

    5. 5

      Publish the offer

      An Enterprise page section and a datasheet.

      Founder/CEO 1 week

    6. 6

      Review

      Deal cycle, pricing and renewals.

      Founder/CEO quarterly

    First result
    2 quarters
    Cost
    Founder sales time and engineering for offline licensing.
    Owner
    Founder/CEO

    Smallest test

    Send the one-page offer to two current design partners.

    Why it works

    Define one productised offer: an annual on-prem Pyrite licence with an offline mode, a validation pack for the customer's chips, an onboarding sprint to wire it into their CI, and a support SLA. Sign three design partners at a discount in exchange for case studies, then publish the offer.

    Regulated teams pay for local deployment and evidence, and a licence turns individual pilots into larger annual contracts.

    Risks and how to handle them

    • Custom work per customer

      Keep the offer fixed and charge for extras.

    • Long procurement

      Start with a paid pilot inside an existing budget.

    Builds on

    Evidence

    • Y CombinatorCompany profile: Fall 2026 batch, founded 2025, team of 4, Waterloo, Canada; products are Cloud Sim (a cloud virtual dev board), Simantic Bot (an AI reviewer that runs firmware on a simulated microcontroller) and Pyrite, "a firmware simulation model available to select partners".Source
    • Y CombinatorLaunch post: "We built the world's first agent-first firmware simulator"; Cloud Sim is "the first virtual devboard you can program and run entirely from your terminal" and needs only an MCU model and a compiled ELF; Pyrite can run locally in CI/CD or agent harnesses for select partners; the founders, formerly in firmware and hardware roles at Tesla, Neuralink, NXP and Onsemi, ask to hear from teams in automotive, robotics or aerospace. No traction or customers are disclosed.Source
    • Your websitePricing page: Community ("For individuals getting started", no price shown, "$10 in monthly credits", 3 concurrent simulation runs, GitHub integration, "Sign up"); Developer ("For professional engineers", $100 a month or $1,000 a year, shown as $83.33/mo, "$200 in monthly credits", 10 concurrent runs, early access to simulation models); Enterprise ("Contact Us": unlimited API usage, unlimited concurrent runs and PR reviews, private hosting, custom subdomain and SSO, custom SLA). There is no FAQ and no explanation of what a credit buys.Source
    • Your websiteValidation page: principles such as "Let real hardware be the reference", "Look beyond a successful boot", "Make failures part of the test" and "Show the evidence. Name the boundaries.", and "We are building toward a validation standard that firmware simulation has not seen before." It names no chips and shows no numbers or results; the call to action is "Talk to the team".Source

Wildcards 2

Unexpected experiments, each with a cheap way to test it.

  • 22

    Real visitors already pass Core Web Vitals (field LCP about 1.1 s), so this is polish, not a problem. In Google's lab phone test, though, the hero circuit-board image is the largest element and paints at 2.9 s. Images are 1,458 KB of a 1,555 KB page, and render-blocking font requests cost about 1.2 s. As traffic shifts to mobile links shared from social and launch posts, a lighter hero keeps first impressions crisp.

    Low impactLow effort First result in 1 day Founder/CTO

    First action

    Re-export the hero board image as AVIF at its rendered size.

    How to measure it

    Mobile lab LCP

    Now
    –
    Target
    under 2.0 s
    By
    2 weeks

    Steps

    1. 1

      Compress the hero

      AVIF/WebP at rendered sizes.

      Founder/CTO 1 hour

    2. 2

      Prioritise it

      fetchpriority=high on the hero image.

      Founder/CTO 15 minutes

    3. 3

      Self-host fonts

      With font-display swap.

      Founder/CTO 1 hour

    4. 4

      Preload the main font

      One weight only.

      Founder/CTO 15 minutes

    5. 5

      Fix contrast

      The one low-contrast text flagged.

      Founder/CTO 30 minutes

    6. 6

      Retest

      PageSpeed mobile and desktop.

      Founder/CTO 15 minutes

    First result
    1 day
    Cost
    $0
    Owner
    Founder/CTO

    Smallest test

    Swap the hero image format and rerun PageSpeed.

    Why it works

    Serve the hero board image as AVIF/WebP at the displayed sizes with a high fetch priority, self-host the fonts with font-display swap, and preload the main font. Recheck PageSpeed afterwards.

    A faster first paint on slower phones protects the first impression for launch-driven traffic at almost no cost.

    Risks and how to handle them

    • Image quality loss on the hero

      Compare at 2x density before shipping.

    Evidence

    • PageSpeedGoogle PageSpeed (homepage, tested this week): mobile score 92, desktop 100; real visitors pass Core Web Vitals (field LCP 1.06 s, CLS 0); in the lab phone test the largest element, the hero circuit-board image, paints at 2.9 s; images are 1,458 KB of a 1,555 KB page; render-blocking requests (Google Fonts) cost about 1.2 s; one colour-contrast accessibility issue.Source
  • 23

    Reviews are private by default but can be shared with anyone who has the link. The homepage has no customers or testimonials, and the GitHub org's repos have no stars yet. Every passing simulated review in a public repository is proof that other firmware engineers would see, if it were visible.

    Medium impactLow effort First result in 1 quarter Founder/CTO

    First action

    Add a "Get badge" button to the review page for public repositories.

    How to measure it

    Badge-referred signups

    Now
    –
    Target
    50
    By
    2 quarters

    Steps

    1. 1

      Design the badge

      Chip, status and date.

      Founder/CTO half a day

    2. 2

      Build the summary page

      Public, minimal, owner-controlled.

      Founder/CTO 3 days

    3. 3

      Add the button

      On review and CI pages for public repos.

      Founder/CTO 1 day

    4. 4

      Seed it

      Student teams and open-source maintainers first.

      Founder/CEO 1 month

    5. 5

      Add attribution

      Track signups from badge pages.

      Founder/CTO 2 hours

    6. 6

      Measure

      Badges live and referred signups.

      Founder/CEO monthly

    First result
    1 quarter
    Cost
    About a week of engineering.
    Owner
    Founder/CTO

    Smallest test

    Hand-make badges for two friendly open-source projects.

    Why it works

    Let users add a README badge ("Simulated on STM32F401RE: passing") that links to a public summary of the latest review or CI run, with a small "Try Simantic on your firmware" link. Make it opt-in per repository, and show only the summary unless the owner shares more.

    Badges are a familiar open-source growth loop, and each one is a credible endorsement from a working engineer in place of the logos Simantic doesn't yet have.

    Risks and how to handle them

    • Exposing private details

      Opt-in only, public repos only, summary by default.

    Builds on

    Evidence

    • Your websitePR reviewer docs (/docs/pr-review): mention @simantic-bot in a pull request and it builds the firmware, runs scenarios designed from the diff in the simulator and posts one review; it is "an on-demand assistant, not CI: nothing runs on a push"; "a pull request for any other MCU gets a static review, without simulation and never approved"; reviews draw on "$10 a month on the Community plan, $200 on Developer", shared with simulation runs, and are refused with less than $1.50 left; enterprise groups are billed as a group; reviews are private by default and can be shared by link; the diff, build output and simulator output are sent to the company's model provider.Source
    • Your websiteHomepage: H1 "Real firmware. Simulated hardware."; Cloud Sim access is listed as HTTP API, TypeScript and MCP, with "Python" and "CLI" marked "Coming soon"; the AI firmware review section shows a sample review of an STM32F072RB blinky project; the hero visual shows I2C and RC rise-time waveforms; "Backed by Y Combinator"; no customers, logos or testimonials, and no list of supported microcontrollers.Source
    • GitHubThe simantic-dev organization has 17 public repositories, all at 0 stars: `pippy` ("Test firmware without a board: Python SDK and pytest plugin", alpha 0.4.x, "not distributed on PyPI"); `embedded-ci-action` (a GitHub Action that submits firmware .elf files for simulation, with results in the dashboard's Embedded CI tab; no releases, used at @main); `examples` (good and faulty scripted peers per bus: GPS and RC links over UART, a barometer, power monitor and accelerometer over I2C, an IMU over SPI, DroneCAN and Ethernet; "No firmware is included"); and forks of Zephyr, QEMU and open-source flight-controller, e-reader, solar-car and student racing firmware.Source
More example playbooks and reviews
ProgrammableDigital

Show Simantic’s product working

Turn screenshots or the website into an animated product demo in its own brand, for the homepage, launch posts and outreach in this playbook.

Make a demo (opens programmable.digital in a new tab)

More growth playbooks

Other companies we have researched, each read from what they are hiring for.

  • Polymarket growth strategyA peer-to-peer prediction market where people trade contracts on real-world events, run as two separate products: polymarket.com, the…21 growth ideas
  • Tide growth strategyA business finance platform for small businesses: business accounts, cards, payments, invoicing and accounting, company formation, lending…21 growth ideas
  • HubSpot growth strategyAn AI-powered customer platform (Marketing, Sales, Service, Content, Data and Revenue Hubs on Smart CRM) with Agent Hub for building and…21 growth ideas
  • Tavus growth strategyReal-time conversational video AI: lifelike AI humans ("PALs") that see, hear and respond face to face, sold as a developer API (CVI), a…21 growth ideas
  • UN/COMMON growth strategyA Shopify Plus and Klaviyo agency offering lifecycle (email and SMS) marketing, redesign and replatforming, conversion optimization, and…21 growth ideas
  • Knak growth strategyAn enterprise marketing production platform, now positioned as the agentic production layer: teams and AI agents build on-brand emails…21 growth ideas
  • Jiminy's growth strategyDog food, treats and dental chews made with insect protein (crickets and black soldier fly larvae), sold online and through pet stores…21 growth ideas
  • DataDome growth strategyBot and AI agent trust management: real-time detection of bots, fraud and AI agents across websites, apps, APIs and MCP servers, with…20 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on November 10, 2026, and cites them throughout. Simantic did not commission, review or endorse it, and Programmable is not affiliated with Simantic. Company and product names are trademarks of their owners, used only to identify them. The ideas are suggestions and opinion, not professional, legal or financial advice, and not statements about how Simantic runs its business.

Something wrong, or do you represent Simantic? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

What to focus on in each month, and the ideas that carry it.

  1. Days 1 to 307 ideas

    Make the product legible: public chip list, real Cloud Sim quickstart, credit math, correct "coming soon" labels, a CI page, a readable careers page and a smarter contact form.

  2. Days 31 to 6011 ideas

    Build proof and loops: validation results, a model changelog, public bug finds, student teams, the PyPI SDK, comparison pages, MCP listings, sample firmware, one-click config, review badges and a lighter hero.

  3. Days 61 to 905 ideas

    Open the bigger revenue: an enterprise data page, a Team plan, board-level electronics simulation, chip-maker partnerships and a local Pyrite licence.

The research behind the plan

3 research areas, each summarised in a few lines. Open any of them for the full detail.

Customers

What customers say

Customers value running real firmware without flashing a physical board. The most cited problem: waiting weeks or months for physical devices.

  • Deterministic runs: the same ELF and inputs give the same output
  • Slow, expensive hardware-in-the-loop setups
  • In their words: “real firmware”

Market

Competitive landscape

3 competitors compared: Renode (Antmicro), Wokwi and Embedder.

  • Gap to exploit: Published silicon-versus-simulation accuracy per chip
  • Renode (Antmicro): Services-led; Simantic's founders cite speed and timing accuracy as reasons not to build on it
  • Wokwi: Maker and education focus, not regulated industry

Hiring signals

Simantic has one open role, and it shows where the company is going: a Lead Simulation Researcher ($200K-$300K, 2-5% equity, Waterloo or San Francisco) to build an electronics simulation platform on SPICE and Xyce for…

  • Lead Simulation Researcher (Electronics / Firmware): SPICE/Xyce electronics simulation for agents, validated on hardware. Ideas 6, 8, 19.

From the team that wrote this playbook

All products
  • ProgrammableAgencyWorking AI systems in days, not quarters (opens programmable.agency in a new tab)
  • ProgrammableInfoAn agent that finds, researches and asks first (opens programmable.info in a new tab)
  • ProgrammableDigitalScreenshots in, an animated product demo out (opens programmable.digital in a new tab)
  • ProgrammableWatchKnow the day a supplier falls out of compliance (opens programmable.watch in a new tab)
  • ProgrammableTalkGet registrants for your events, and know who’s coming (opens programmable.talk in a new tab)
  • ProgrammablePageText what you need, your customer gets a page that works (opens programmable.page in a new tab)
  • ProgrammableCommunityForums with labelled AI members, never empty (opens programmable.community in a new tab)
ProgrammableMarketing
ExamplesLeaderboardsProductsTermsPrivacyCookiesDPALegalContact

© 2026 AI-Powered Growth Intelligence