ProgrammableMarketing/cua.ai (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 6
  • This quarter · 12
  • Big bets · 2
  • Wildcards · 2
  • How we can help
Generated Nov 10, 2026 1 view
  1. Programmable
  2. Playbooks
  3. Cua
Cua logo

growth playbook · Nov 10, 2026

cua.ai (opens in a new tab)

Cua growth playbook

Diagnosis

Cua has unusual developer pull for a three-person YC company: 32.1k GitHub stars, use in projects from 15 AI labs and developer platforms, a weekly changelog and a rigorous benchmarks library. It now runs two businesses on one site.

Self-serve Spaces plans are well explained, but the sales-led side its new GTM lead is hired to grow (environments, trajectory data, Driver deployments for AI labs) has no front door: the Environments Catalog is unlinked from the homepage, pricing page and footer, the pricing page has no enterprise or contact option, and help goes to one person's inbox.

This playbook opens that door first, fixes the sitemap and stale recruiting copy, then turns the comparison and benchmark work into switching pages, proof stories and an enterprise plan, and finishes with bigger bets on private evaluations and partner clouds.

Growth Score

7/10

Biggest Opportunity

Turn open-source adoption by AI labs and agent teams into sales-led revenue by giving the Environments Catalog, Driver deployments and larger teams a visible, routed path to a conversation.

Biggest Risk

Hosted sandbox clouds such as E2B and Daytona already sell compliance, enterprise pages and named proof, and a three-person team can spread itself across too many products before any one has a repeatable sales motion.

01

Start here

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

  1. 1

    Give AI labs a front door: link the Environments Catalog and a "talk to us" plan from the homepage and pricing page

    Add a "Labs and enterprise: talk to us" column to /pricing that links to /data and a short form.

    High impact First result in 2 weeks
  2. 2

    Put the pricing, support, download and bring-your-own-computers pages in the sitemap

    Add the missing routes to the sitemap generator.

    Medium impact First result in 2 weeks
  3. 3

    Replace the single personal inbox with routed sales and support contact, and wire Cua-Bench's "get in touch" to it

    Create sales@, support@ and security@ on cua.ai and forward them to a shared inbox.

    Medium impact First result in 1 week

02

Your idea portfolio

22 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 · 1Big bets · 10Fill-ins · 8Reconsider · 3
Effort →

Showing 22 of 22 ideas

This week 6

Quick wins with a clear first action.

  • 01

    Cua's highest-value offer is hard to find. The Environments Catalog at cua.ai/data promises "4K+ environments and 400K+ tasks", trajectory data and SOC 2 Type I, but nothing on the homepage, the pricing page or the footer links to it (it is reachable only from the sitemap). The pricing page stops at Teams, "1 to 100 seats", with no enterprise tier, no contact-sales option and no email, and the site history shows an Enterprise plan sat next to Pro a year ago. Meanwhile the new Founding Technical GTM Lead is hired to sell evaluation environments, trajectory data and Driver deployments to AI labs and agent companies. A lab engineer who arrives from the 32k-star repo sees $15 and $30 plans and leaves.

    High impactLow effort First result in 2 weeks Founding Technical GTM Lead

    First action

    Add a "Labs and enterprise: talk to us" column to /pricing that links to /data and a short form.

    How to measure it

    Inbound lab and enterprise conversations

    Now
    –
    Target
    10 a month
    By
    60 days

    Steps

    1. 1

      Write the column

      What labs and larger teams can buy: environments, trajectory data, Driver deployments, 100+ seats, with one call to action.

      Founding Technical GTM Lead 2 hours

    2. 2

      Build one form

      Company, use case, volume, timeline; send to the CRM and a shared channel, not a personal inbox.

      Founding Engineer, Infra & Agent Systems 1 day

    3. 3

      Add it to /pricing

      A fourth column beside Free, Pro and Teams.

      Founding Engineer, Infra & Agent Systems 2 hours

    4. 4

      Link /data everywhere it belongs

      Homepage nav, footer Products list and a "For labs" line under the projects strip.

      Founding Engineer, Infra & Agent Systems 2 hours

    5. 5

      Swap the email links

      "Request the full catalog", "Request scope" and the Cua-Bench prompt open the same form.

      Founding Engineer, Infra & Agent Systems 2 hours

    6. 6

      Show the trust signal

      SOC 2 Type I and the licences next to the form.

      CEO 1 hour

    7. 7

      Answer within a working day

      Reply with the sample pack and two time slots.

      Founding Technical GTM Lead ongoing

    8. 8

      Review monthly

      Conversations, use case mix and deals started; adjust the column wording.

      Founding Technical GTM Lead monthly

    First result
    2 weeks
    Cost
    A day of engineering time
    Owner
    Founding Technical GTM Lead

    Smallest test

    Add the pricing column and the nav link today and count form submissions for two weeks.

    Why it works

    Add one visible path for buyers who will never use a credit card. (1) A fourth column on /pricing: "Labs and enterprise: environments, trajectory data, Driver deployments, more than 100 seats, SSO when it ships. Talk to us." (2) An "Environments" item in the homepage nav and a "For labs" line under the "Used in projects from" strip, both pointing to /data. (3) /data, /cuabench and the new pricing column all use one short form (company, use case: evals, training data, fleet or Driver; volume; timeline) that lands in the GTM lead's pipeline, instead of an email link. (4) Put "SOC 2 Type I" next to the form and add /data to the footer under Products. Measure how many qualified conversations it creates per month.

    Developer-first companies usually convert their biggest accounts through a contact path that sits next to self-serve pricing. Cua already has the hard parts: the catalog, the benchmark credibility, the logos and a compliance report. Making them visible costs a few hours and gives the new GTM lead inbound demand from day one instead of only cold outreach.

    Ready to use

    ChecklistLabs and enterprise path launch list

    [ ] /pricing: fourth column "Labs and enterprise" with Talk to us
    [ ] Homepage nav: Environments -> /data
    [ ] Footer Products: Environments Catalog
    [ ] "For labs" line under Used in projects from
    [ ] One form on /pricing, /data and /cuabench (company, use case, volume, timeline)
    [ ] Form goes to CRM + shared channel
    [ ] SOC 2 Type I and licences next to the form
    [ ] Auto-reply with the sample pack
    [ ] Monthly count of qualified conversations

    Risks and how to handle them

    • Low-quality inbound from hobbyists

      Ask for use case and volume, and route small requests to Pro or Teams automatically.

    • A tiny team cannot answer fast

      Auto-reply with the sample pack and a booking link until the GTM lead starts.

    Evidence

    • Your websiteEnvironments Catalog (/data): "Train and evaluate agents on real computer-use tasks."; "4K+ environments and 400K+ tasks" on Linux, Windows and macOS; scoped data packages with trajectories, annotations and human-reviewed golden runs; a five-environment sample pack; "SOC 2 Type I"; no prices; "Request the full catalog" and "Request scope" open an email to a contact address on the older trycua.com domain. The page is in the sitemap but not linked from the homepage, the pricing page or the footer.Source
    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
    • Your websiteHomepage (checked November 10, 2026): headline "Combine all your computers into one supercomputer."; plan limits (Free: 2 computers in use at once and 20 remote hours a month; Pro: 5 and 200; each Teams seat adds 5 and 200); a "Used in projects from" strip of 15 AI labs, developer platforms and agent projects, each linking to a public code file or doc rather than a story; no testimonials; no link to /data, "enterprise", "contact sales", "SOC 2" or a demo.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
    • Site historyChanges since November 2025: starting price $10 to $15 a month; plans then "Pro | Enterprise", now "Free | Pro | Teams"; headline moved from "Computer Use Agent Platform" to "Combine all your computers into one supercomputer."Source
  • 02

    Cua's sitemap lists 320 URLs, including 227 docs pages and 18 benchmarks, but not /pricing, /support, /bring-your-own-computers, /careers or the /download pages. The pages that turn a visitor into a user or a customer are the ones missing from the file that tells search engines and AI crawlers what the site contains.

    Medium impactLow effort First result in 2 weeks Founding Engineer

    First action

    Add the missing routes to the sitemap generator.

    How to measure it

    Organic sessions on pricing and download pages

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      List public routes

      Every marketing page outside docs, blog and benchmarks.

      Founding Engineer, Infra & Agent Systems 1 hour

    2. 2

      Fix the generator

      Include static routes with lastmod.

      Founding Engineer, Infra & Agent Systems 2 hours

    3. 3

      Link the sandboxes comparison

      From the /compare hub.

      Founding Engineer, Infra & Agent Systems 30 minutes

    4. 4

      Resubmit

      Search Console and Bing Webmaster Tools.

      Founding Engineer, Infra & Agent Systems 30 minutes

    5. 5

      Check queries

      Brand plus pricing and download searches.

      Founding Technical GTM Lead 2 weeks

    6. 6

      Add a test

      Fail the build if a marketing route is missing.

      Founding Engineer, Infra & Agent Systems 2 hours

    First result
    2 weeks
    Cost
    $0
    Owner
    Founding Engineer, Infra & Agent Systems

    Smallest test

    Add /pricing today and watch its crawl date in Search Console.

    Why it works

    Generate the sitemap from the site's routes rather than from content collections, so every public marketing page is included, with lastmod dates. Add /pricing, /bring-your-own-computers, /support, /download/mac, /download/windows, /download/linux and /compare/sandboxes links in the hub. Resubmit in Search Console and check that "cua pricing" and "cua spaces download" queries land on the right pages.

    Search engines can find linked pages without a sitemap, but listing them with fresh dates speeds up recrawling after every price or plan change, and it is the cheapest technical fix on the site.

    Risks and how to handle them

    • Small measurable effect

      Treat it as hygiene; it takes an afternoon.

    Evidence

    • Your websiteSitemap (one urlset, 320 URLs, checked November 10, 2026): 227 docs pages, 31 blog posts, 18 benchmarks, /compare, /compare/clouds, /compare/sandboxes, /data and /cuabench; it does not list /pricing, /support, /bring-your-own-computers, /careers or the /download pages; there are no /customers, /case-studies, /enterprise, /integrations, /use-cases or /alternatives pages.Source
    • Your website"Cua Spaces and hosted sandboxes compared" (checked October 24, 2026): E2B, Daytona, Modal Sandboxes, Vercel and Cloudflare side by side on price per sandbox-hour (Cua "$0 on your own machines" against $0.0666 to $0.1492 for 1 vCPU and 1 GiB), longest run, concurrency (448 sandboxes on one Mac Studio), cold start and licences; Daytona, Vercel and Cloudflare cold starts are still "To be measured". The /compare hub itself links only to /compare/clouds and the benchmarks.Source
  • 03

    The support page sends billing, plan, account and security questions to one named person's mailbox, offers no sales contact, and Cua-Bench's "Interested in ... get in touch" prompt has nothing behind it. Catalog requests go to a contact address on the older trycua.com domain. Buyers, users and security reporters all compete for one person's attention.

    Medium impactLow effort First result in 1 week CEO

    First action

    Create sales@, support@ and security@ on cua.ai and forward them to a shared inbox.

    How to measure it

    First response to sales and support email

    Now
    –
    Target
    Under 1 working day
    By
    30 days

    Steps

    1. 1

      Create addresses

      sales, support and security on cua.ai.

      CEO 1 hour

    2. 2

      Pick a shared inbox

      A helpdesk with tags for billing, plan and bugs.

      CEO 2 hours

    3. 3

      Update /support

      Show the addresses and response times.

      Founding Engineer, Infra & Agent Systems 1 hour

    4. 4

      Wire Cua-Bench

      "Get in touch" opens the sales form.

      Founding Engineer, Infra & Agent Systems 1 hour

    5. 5

      Move catalog requests

      From the trycua.com address to the form.

      Founding Engineer, Infra & Agent Systems 1 hour

    6. 6

      Review weekly

      Which questions recur; turn them into docs.

      Founding Technical GTM Lead weekly

    First result
    1 week
    Cost
    Helpdesk seat costs
    Owner
    CEO

    Smallest test

    Put a sales address on /support this week and count messages.

    Why it works

    Create role addresses on cua.ai (sales, support, security), route them to a shared helpdesk and to the CRM, and show them on /support with expected response times. Give security its own address and a short disclosure note on /security-research. Point the Cua-Bench prompt at the sales form from idea 1.

    Role addresses survive hiring and holidays, let the GTM lead see every buying signal, and look like a company that can support an enterprise account.

    Risks and how to handle them

    • Shared inbox nobody owns

      Name an owner per tag and review in the weekly meeting.

    Builds on

    Evidence

    • Your websiteSupport page: one named person's mailbox at trycua.com "for billing, plan, account and security questions", plus docs, Discord and GitHub issues; no sales or enterprise contact.Source
    • Your websiteCua-Bench: "Make your agents better at computers."; 25 expert-authored KiCad schematic tasks on which the best frontier agent clears 6 of 25; a joint report with Snorkel AI; an "Interested in" prompt labelled "get in touch" with no form or email behind it.Source
    • Your websiteEnvironments Catalog (/data): "Train and evaluate agents on real computer-use tasks."; "4K+ environments and 400K+ tasks" on Linux, Windows and macOS; scoped data packages with trajectories, annotations and human-reviewed golden runs; a five-environment sample pack; "SOC 2 Type I"; no prices; "Request the full catalog" and "Request scope" open an email to a contact address on the older trycua.com domain. The page is in the sitemap but not linked from the homepage, the pricing page or the footer.Source
  • 04

    cua.ai/careers redirects to the Y Combinator jobs page, which still says "Give every agent a cloud desktop" while the site now says Cua does not host, rent or resell computers. The GTM post cites "9k+ GitHub stars in just 4 months" when the repo has 32.1k, the engineer post calls Cua "a cloud container platform", and "Research Intern (Summer 2026)" is still open in November. Candidates, investors and the labs the GTM lead will call all read these pages.

    Medium impactLow effort First result in 2 weeks CEO

    First action

    Change the star count and tagline on the YC page.

    How to measure it

    Qualified GTM applicants

    Now
    –
    Target
    +50%
    By
    45 days

    Steps

    1. 1

      Update the tagline

      Match the homepage message.

      CEO 15 minutes

    2. 2

      Update traction

      Stars, forks, projects and benchmarks.

      CEO 30 minutes

    3. 3

      Rewrite product lines

      Spaces on owned hardware, Driver, catalog.

      CTO 1 hour

    4. 4

      Close or re-date the intern post

      Summer 2027 if still wanted.

      CTO 15 minutes

    5. 5

      Add /careers on cua.ai

      Roles plus links to YC, and add it to the sitemap.

      Founding Engineer, Infra & Agent Systems 2 hours

    6. 6

      Set a quarterly check

      Review every public profile each quarter.

      CEO quarterly

    First result
    2 weeks
    Cost
    $0
    Owner
    CEO

    Smallest test

    Fix the GTM post's traction line today.

    Why it works

    Rewrite the YC company tagline and the three posts around Spaces on owned hardware, Cua Driver and the Environments Catalog; update traction (32.1k stars, the 15 projects strip, the benchmarks library); close or re-date the intern post; and add a short careers page on cua.ai that lists the roles and links to YC, so the footer link stays on the brand domain.

    A founding GTM hire is a sales problem in itself: the best candidates judge momentum from the post, and stale numbers undersell it by more than three times.

    Risks and how to handle them

    • Profiles drift again after the next pivot

      Keep one source-of-truth company blurb and paste it everywhere.

    Evidence

    • HiringCua's Y Combinator jobs page (cua.ai/careers redirects there): batch P25, team size 3, tagline "Give every agent a cloud desktop."; three open roles: Founding Technical GTM Lead ($150K-$200K, 0.50%-1.50% equity), Founding Engineer, Infra & Agent Systems ($100K-$150K) and "Research Intern (Summer 2026)", still listed in November 2026.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
    • HiringFounding Engineer, Infra & Agent Systems: "Design and scale infrastructure to run real AI agents on isolated macOS and Linux containers", plus multi-cloud orchestration and developer-facing APIs and CLIs; the post describes Cua as "A cloud container platform for sandboxed, scalable agent execution environments" used by "thousands of developers".Source
    • GitHubtrycua/cua: 32.1k stars and 2.4k forks; the README leads with Cua Spaces and lists Pro at $18 a month and Teams at $30 a seat (the monthly prices, while the site leads with $15 and $25 billed yearly); a "Deprecated packages" section points older users to a migration guide; MIT for most of the repo, Spaces under FSL-1.1-MIT.Source
    • Your website"Bring the computers. Spaces does the rest.": Cua does not host, rent or resell computers; Spaces runs on your Mac, your other computers or your own AWS, Google Cloud or Modal accounts; the relay is free up to 2 computers in use and 20 remote hours a month; no contact path.Source
  • 05

    For most developers the 32.1k-star repo is the first page they read. Its README lists Pro at $18 a month and Teams at $30 a seat, the monthly prices, while cua.ai leads with $15 and $25 billed yearly. A visitor who compares the two sees different numbers for the same plan, and the README's main button points to a separate spaces.cua.ai address rather than the pricing page.

    Medium impactLow effort First result in 1 week CTO

    First action

    Edit the README pricing lines to match the site.

    How to measure it

    GitHub-referred signups

    Now
    –
    Target
    Measured weekly, +15%
    By
    60 days

    Steps

    1. 1

      Match the prices

      Yearly and monthly, as on /pricing.

      CTO 30 minutes

    2. 2

      State the free limits

      2 computers, 20 remote hours.

      CTO 15 minutes

    3. 3

      Link to pricing

      And to the labs path.

      CTO 15 minutes

    4. 4

      Tag links

      UTM source=github on every cua.ai link.

      Founding Engineer, Infra & Agent Systems 1 hour

    5. 5

      Check the main button

      Make sure it lands on a page in analytics.

      Founding Engineer, Infra & Agent Systems 30 minutes

    6. 6

      Report

      GitHub-referred signups weekly.

      Founding Technical GTM Lead weekly

    First result
    1 week
    Cost
    $0
    Owner
    CTO

    Smallest test

    Fix the two price lines today.

    Why it works

    Replace the README price lines with a small table that matches /pricing (yearly and monthly), states the free limits (2 computers, 20 remote hours), and links to cua.ai/pricing and the labs path from idea 1. Add UTM tags to every README link so the GTM lead can see how much pipeline the repo creates.

    Consistent prices remove a moment of doubt, and tagged links finally measure the repo as the top-of-funnel it already is.

    Risks and how to handle them

    • Prices change again and drift

      Link to /pricing instead of repeating numbers if they change often.

    Evidence

    • GitHubtrycua/cua: 32.1k stars and 2.4k forks; the README leads with Cua Spaces and lists Pro at $18 a month and Teams at $30 a seat (the monthly prices, while the site leads with $15 and $25 billed yearly); a "Deprecated packages" section points older users to a migration guide; MIT for most of the repo, Spaces under FSL-1.1-MIT.Source
    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
  • 16

    Real visitors already pass Core Web Vitals (LCP 1.75 s, INP 96 ms), so this is not urgent. But in Google's phone test the homepage scores 78 with a 5.1 s LCP, and the LCP element is the hero illustration; PageSpeed estimates 225 KB of image savings, 106 KB of unused JavaScript and 0.35 s of render-blocking requests.

    Low impactLow effort First result in 1 day Founding Engineer

    First action

    Re-export the hero image at mobile size and mark it high priority.

    How to measure it

    PageSpeed mobile score

    Now
    –
    Target
    90+
    By
    30 days

    Steps

    1. 1

      Resize the hero

      AVIF/WebP at mobile widths.

      Founding Engineer, Infra & Agent Systems 1 hour

    2. 2

      Prioritise it

      fetchpriority=high and preload.

      Founding Engineer, Infra & Agent Systems 30 minutes

    3. 3

      Trim JavaScript

      Defer the 106 KB not used on load.

      Founding Engineer, Infra & Agent Systems 2 hours

    4. 4

      Fix caching

      Long cache lifetimes on static files.

      Founding Engineer, Infra & Agent Systems 1 hour

    5. 5

      Re-test

      PageSpeed mobile.

      Founding Engineer, Infra & Agent Systems 30 minutes

    6. 6

      Watch field data

      Keep Core Web Vitals passing.

      Founding Engineer, Infra & Agent Systems monthly

    First result
    1 day
    Cost
    $0
    Owner
    Founding Engineer, Infra & Agent Systems

    Smallest test

    Swap the hero image and re-run PageSpeed.

    Why it works

    Serve the hero illustration as a properly sized AVIF or WebP with fetchpriority=high, defer the non-critical script, and set long cache lifetimes on static assets. Re-test on PageSpeed and keep the field numbers passing.

    Developers often open the site on a phone from a link on X or Discord; a faster first view protects that moment for an afternoon's work.

    Risks and how to handle them

    • Effort spent where users are already fine

      Cap it at one afternoon.

    Evidence

    • Google PageSpeedHomepage, November 10, 2026: mobile score 78, desktop 99; lab LCP 5.1 s on mobile (the hero illustration image), but real Chrome visitors pass Core Web Vitals (LCP 1.75 s, INP 96 ms, CLS 0) and LCP has improved 32%; suggested savings: render-blocking requests 0.35 s, 106 KB unused JavaScript, 225 KB images, 304 KB from cache lifetimes.Source

This quarter 12

The core plays for the next three months.

  • 06

    The homepage shows 15 AI labs, developer platforms and agent projects that use Cua, but each logo links to a code file or a doc page, and there are no testimonials or case studies anywhere on the site. The GTM lead's post asks for "referenceable customers with measurable outcomes" and "case studies"; rivals such as Daytona show named customer quotes on their homepage.

    High impactMedium effort First result in 6 weeks Founding Technical GTM Lead

    First action

    Email the maintainers of three listed projects asking for a short interview.

    How to measure it

    Proof stories live

    Now
    –
    Target
    3
    By
    90 days

    Steps

    1. 1

      Choose three

      Projects with a measurable outcome and an active maintainer.

      Founding Technical GTM Lead 1 day

    2. 2

      Ask

      Short interview and approval of the final text.

      Founding Technical GTM Lead 1 week

    3. 3

      Interview

      Problem, choice, setup, result.

      Founding Technical GTM Lead 2 weeks

    4. 4

      Write

      600 words and one number each.

      CEO 1 week

    5. 5

      Publish

      A /customers section, added to the sitemap.

      Founding Engineer, Infra & Agent Systems 2 days

    6. 6

      Reuse

      Pull quotes on the homepage, pricing and outreach.

      Founding Technical GTM Lead ongoing

    First result
    6 weeks
    Cost
    Staff time only
    Owner
    Founding Technical GTM Lead

    Smallest test

    Publish one story from a research project first.

    Why it works

    Pick the three projects with the clearest outcome (for example an evaluation harness, an agent product and a research benchmark), ask their maintainers for a 20-minute interview, and publish short stories: the problem, why Cua Driver or Spaces, the setup, and a number (tasks run, time saved, platforms covered). Link each logo to its story where one exists, and add one pull quote to the homepage and pricing page.

    Logos say "someone uses this"; a story says "a team like yours got this result", which is what a lab buyer needs to start an evaluation.

    Risks and how to handle them

    • Large organisations will not approve quotes

      Start with open-source maintainers and research groups, who can usually speak for their own project.

    Evidence

    • Your websiteHomepage (checked November 10, 2026): headline "Combine all your computers into one supercomputer."; plan limits (Free: 2 computers in use at once and 20 remote hours a month; Pro: 5 and 200; each Teams seat adds 5 and 200); a "Used in projects from" strip of 15 AI labs, developer platforms and agent projects, each linking to a public code file or doc rather than a story; no testimonials; no link to /data, "enterprise", "contact sales", "SOC 2" or a demo.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
    • Competitor researchDaytona's homepage leads with "Talk to Our Team", shows customer logos and two named customer testimonials, and sells customer-managed compute with HIPAA, SOC 2 and GDPR.Source
  • 07

    Cua already wrote a careful comparison against E2B, Daytona, Modal, Vercel and Cloudflare, with per-hour prices and measured cold starts, but it lives on one page that the /compare hub does not even link to, and three rivals' cold starts are still "To be measured". People search "E2B alternative" or "Daytona vs ..." one rival at a time, and the site has no /alternatives or per-rival pages.

    High impactMedium effort First result in 1 month Founding Technical GTM Lead

    First action

    Publish /compare/e2b from the existing table and cold-start benchmark.

    How to measure it

    Downloads and signups from /compare pages

    Now
    –
    Target
    200 a month
    By
    90 days

    Steps

    1. 1

      Pick the order

      Rivals that come up most in Discord and calls.

      Founding Technical GTM Lead 1 day

    2. 2

      Run missing benchmarks

      Same method, dated and reproducible.

      Founding Engineer, Infra & Agent Systems 1 week

    3. 3

      Write each page

      Rows, benchmark, fair limits, switching.

      Founding Technical GTM Lead 1 week

    4. 4

      Check facts

      Rivals' current published prices and limits.

      Founding Technical GTM Lead 2 days

    5. 5

      Link and index

      Hub, sitemap and docs.

      Founding Engineer, Infra & Agent Systems 1 day

    6. 6

      Refresh quarterly

      Re-check prices and re-run benchmarks.

      Founding Engineer, Infra & Agent Systems quarterly

    First result
    1 month
    Cost
    Staff time and a little cloud spend for benchmarks
    Owner
    Founding Technical GTM Lead

    Smallest test

    Publish the E2B page and watch its queries for a month.

    Why it works

    Publish /compare/e2b, /compare/daytona and /compare/modal, each with the relevant rows from the existing table, the matching benchmark, a "when the hosted option is the better choice" section for fairness, and a switching section (see idea 8). Run the missing cold-start benchmarks with the same published method. Link all of them from the /compare hub and the sitemap.

    Comparison searches are the highest-intent traffic in developer infrastructure, and Cua's reproducible numbers make its pages more credible than typical vendor tables.

    Risks and how to handle them

    • Rivals dispute numbers

      Keep the published methods and invite them to reproduce.

    Builds on

    Evidence

    • Your website"Cua Spaces and hosted sandboxes compared" (checked October 24, 2026): E2B, Daytona, Modal Sandboxes, Vercel and Cloudflare side by side on price per sandbox-hour (Cua "$0 on your own machines" against $0.0666 to $0.1492 for 1 vCPU and 1 GiB), longest run, concurrency (448 sandboxes on one Mac Studio), cold start and licences; Daytona, Vercel and Cloudflare cold starts are still "To be measured". The /compare hub itself links only to /compare/clouds and the benchmarks.Source
    • Your websiteBenchmark "Spaces vs E2B cold start" (measured October 25, 2026): create-to-first-command p50 234.6 ms for a Space on a Mac Studio against 299 ms for E2B over the internet, p99 270.1 ms against 877.8 ms, 30 runs each, with raw CSV and JSON and the commands to reproduce it; the page flags that it is not like-for-like. All 18 benchmark pages are dated October 25, 2026.Source
    • Your websiteSitemap (one urlset, 320 URLs, checked November 10, 2026): 227 docs pages, 31 blog posts, 18 benchmarks, /compare, /compare/clouds, /compare/sandboxes, /data and /cuabench; it does not list /pricing, /support, /bring-your-own-computers, /careers or the /download pages; there are no /customers, /case-studies, /enterprise, /integrations, /use-cases or /alternatives pages.Source
    • Competitor researchE2B calls itself "The AI agent cloud": SOC 2 Type II, a Hobby tier with one-time credits, Pro at $150 a month plus usage, and an Enterprise page covering bring-your-own-cloud deployments.Source
  • 08

    The November 2 changelog shipped "compatibility gateways for E2B and Daytona", which means a team could point its existing sandbox code at Spaces. That is a switching tool, but it appears only as one line in a changelog entry, with no guide, no landing page and no cost comparison for the team considering the move.

    High impactMedium effort First result in 1 month Founding Engineer

    First action

    Write the E2B migration guide with a working example repo.

    How to measure it

    Active gateway users

    Now
    –
    Target
    50 teams
    By
    90 days

    Steps

    1. 1

      Document limits

      Which calls are supported today.

      Founding Engineer, Infra & Agent Systems 2 days

    2. 2

      Build examples

      One repo per gateway, before and after.

      Founding Engineer, Infra & Agent Systems 3 days

    3. 3

      Write the guides

      One-line change, test plan, cost example.

      Founding Engineer, Infra & Agent Systems 3 days

    4. 4

      Publish

      Docs plus the per-rival pages.

      Founding Technical GTM Lead 1 day

    5. 5

      Instrument

      Count gateway calls by account.

      Founding Engineer, Infra & Agent Systems 2 days

    6. 6

      Reach out

      Offer help to heavy gateway users.

      Founding Technical GTM Lead ongoing

    First result
    1 month
    Cost
    Engineering time
    Owner
    Founding Engineer, Infra & Agent Systems

    Smallest test

    Publish the E2B guide and count gateway users after two weeks.

    Why it works

    Write a migration guide for each gateway: the one-line change, what works and what does not yet, a cost example using the comparison page's per-hour prices, and a 15-minute test plan. Put the guides on the per-rival pages from idea 7 and in docs, post them in the communities where those SDKs' users ask questions, and offer the GTM lead's help to any team running more than a set number of sandboxes.

    The biggest barrier to switching infrastructure is rewriting code; a gateway removes it, but only if people know it exists and trust it.

    Risks and how to handle them

    • Gaps in compatibility frustrate early switchers

      Publish a supported-calls table and a fast issue path.

    Builds on

    Evidence

    • Your websiteChangelog, November 2, 2026: "compatibility gateways for E2B and Daytona", Hetzner as your own cloud, an elastic pool, "cua share" and a public relay.Source
    • Your website"Cua Spaces and hosted sandboxes compared" (checked October 24, 2026): E2B, Daytona, Modal Sandboxes, Vercel and Cloudflare side by side on price per sandbox-hour (Cua "$0 on your own machines" against $0.0666 to $0.1492 for 1 vCPU and 1 GiB), longest run, concurrency (448 sandboxes on one Mac Studio), cold start and licences; Daytona, Vercel and Cloudflare cold starts are still "To be measured". The /compare hub itself links only to /compare/clouds and the benchmarks.Source
  • 09

    The quickstart takes about five minutes and "A free Cua account is optional": everything on the first computer works without one. That is great for adoption, but it means most new users are anonymous, so Cua cannot send onboarding, announce features or see who is ready for Pro or Teams. The free plan's sharing and remote use only start once an account exists.

    High impactMedium effort First result in 1 month CTO

    First action

    Measure today's share of installs that ever sign in.

    How to measure it

    7-day sign-in rate

    Now
    –
    Target
    +30% relative
    By
    90 days

    Steps

    1. 1

      Baseline

      Installs versus sign-ins, anonymously.

      Founding Engineer, Infra & Agent Systems 2 days

    2. 2

      Design the prompt

      After the first Space, remote access in one click.

      CTO 3 days

    3. 3

      Add the QR code

      Open the Space on a phone.

      Founding Engineer, Infra & Agent Systems 3 days

    4. 4

      Write three emails

      Second computer, teleport, invite.

      Founding Technical GTM Lead 2 days

    5. 5

      Ship to half

      A/B test the prompt.

      Founding Engineer, Infra & Agent Systems 3 weeks

    6. 6

      Report

      Sign-in and Pro conversion by cohort.

      CTO monthly

    First result
    1 month
    Cost
    Engineering time
    Owner
    CTO

    Smallest test

    Add a single sign-in prompt after the first Space for one release.

    Why it works

    After the first Space is created, show one prompt: "Reach this Space from your phone or another computer: sign in free", with a QR code and a one-click link. Follow sign-up with three short emails over two weeks (add a second computer, teleport an app, invite a teammate), each tied to an in-app event, and report the share of installs that sign in.

    Asking at the moment the value of remote access is obvious converts far better than a sign-up wall, and keeps the no-account start that developers like.

    Risks and how to handle them

    • Developers resent nags

      Show the prompt once, dismissible, and never block local use.

    Evidence

    • Your websiteSpaces quickstart: about five minutes, four steps; "A free Cua account is optional" and everything on the first computer works without one; requires macOS 26 or later, Windows (preview) or Ubuntu 24.04; a macOS Space needs an Apple silicon Mac and about 50 GB free.Source
    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
  • 10

    Free allows 2 computers in use at once and 20 remote hours a month; Pro raises that to 5 and 200. When a user goes over, "nothing is cut off" and the app shows a notice with an Upgrade button. That is generous, but a notice at the limit is easy to dismiss, and there is no trial.

    Medium impactLow effort First result in 1 month CTO

    First action

    Count how many free accounts pass 15 remote hours in a month.

    How to measure it

    Monthly free-to-paid upgrades

    Now
    –
    Target
    +25%
    By
    90 days

    Steps

    1. 1

      Size the band

      Accounts near 20 hours or 2 computers.

      Founding Engineer, Infra & Agent Systems 1 day

    2. 2

      Write the notes

      Usage, what Pro adds, one offer.

      Founding Technical GTM Lead 1 day

    3. 3

      Trigger at 75%

      In-app and email.

      Founding Engineer, Infra & Agent Systems 3 days

    4. 4

      Offer at the limit

      A 14-day Pro pass or first month at the yearly rate.

      CTO 2 days

    5. 5

      Test

      Offer versus plain notice.

      Founding Engineer, Infra & Agent Systems 4 weeks

    6. 6

      Review

      Upgrade and churn by band.

      CTO monthly

    First result
    1 month
    Cost
    Engineering time and discount cost
    Owner
    CTO

    Smallest test

    Email this month's heaviest free users their usage and a Pro link.

    Why it works

    Send an in-app and email note at 75% of the remote-hour allowance showing the user's own usage this month, and at the limit offer the first month of Pro at the yearly rate, or a 14-day Pro pass. Tell Teams prospects how many computers their team used together. Track upgrade rate by usage band.

    Users upgrade when they can see their own usage approaching a cap, and a time-boxed offer turns a soft limit into a decision.

    Risks and how to handle them

    • Discounts train users to wait

      Offer the pass once per account.

    Builds on

    Evidence

    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
    • Your websiteHomepage (checked November 10, 2026): headline "Combine all your computers into one supercomputer."; plan limits (Free: 2 computers in use at once and 20 remote hours a month; Pro: 5 and 200; each Teams seat adds 5 and 200); a "Used in projects from" strip of 15 AI labs, developer platforms and agent projects, each linking to a public code file or doc rather than a story; no testimonials; no link to /data, "enterprise", "contact sales", "SOC 2" or a demo.Source
  • 11

    A year ago Cua's plans were Pro and Enterprise; today they are Free, Pro and Teams, with Teams capped at 100 seats and SSO and audit export listed as "coming next". Hosted rivals sell exactly these controls: E2B has an Enterprise page and SOC 2 Type II, and Daytona leads with "Talk to Our Team" and compliance claims.

    High impactMedium effort First result in 1 quarter Founding Technical GTM Lead

    First action

    Write a one-page Enterprise plan sheet and test it on three Teams accounts.

    How to measure it

    Enterprise contracts

    Now
    –
    Target
    2
    By
    2 quarters

    Steps

    1. 1

      List the controls

      SSO, audit export, history, policies.

      CTO 1 day

    2. 2

      Price it

      Per seat with a minimum, annual.

      CEO 2 days

    3. 3

      Build the security pack

      SOC 2 report, licences, data flow.

      CTO 1 week

    4. 4

      Add to pricing

      As the talk-to-us column.

      Founding Engineer, Infra & Agent Systems 1 day

    5. 5

      Pitch largest Teams accounts

      Three conversations.

      Founding Technical GTM Lead 1 month

    6. 6

      Set the roadmap

      Ship the controls the first buyers need.

      CTO 1 quarter

    First result
    1 quarter
    Cost
    Staff time; SSO engineering
    Owner
    Founding Technical GTM Lead

    Smallest test

    Ask the five largest Teams accounts what would make them sign an annual contract.

    Why it works

    Define an annual Enterprise plan: everything in Teams, SSO, audit export and 90-day audit history, more than 100 seats, device approval policies, a security review pack (SOC 2 report, licences, data flow) and a named contact. Price it by seat with a minimum, sell it through the contact path from idea 1, and use the first two contracts to set the roadmap order for SSO and audit export.

    Larger teams buy controls and a contract, not more hours; packaging the controls already on the roadmap creates a higher price point without new infrastructure.

    Risks and how to handle them

    • Selling controls before they ship

      Sell with dated delivery commitments and early pricing.

    Builds on

    Evidence

    • Site historyChanges since November 2025: starting price $10 to $15 a month; plans then "Pro | Enterprise", now "Free | Pro | Teams"; headline moved from "Computer Use Agent Platform" to "Combine all your computers into one supercomputer."Source
    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
    • Competitor researchE2B calls itself "The AI agent cloud": SOC 2 Type II, a Hobby tier with one-time credits, Pro at $150 a month plus usage, and an Enterprise page covering bring-your-own-cloud deployments.Source
    • Competitor researchDaytona's homepage leads with "Talk to Our Team", shows customer logos and two named customer testimonials, and sells customer-managed compute with HIPAA, SOC 2 and GDPR.Source
  • 12

    Cua's trust story is scattered: SOC 2 Type I appears only on the unlinked /data page, teleport approvals and Touch ID sit in the homepage FAQ, telemetry defaults in another FAQ, and licences on the comparison page. Launch coverage singled out teleport, noting that a transferred session gives the Space's controller "a live authenticated foothold". Rivals put SOC 2 and compliance on their homepages.

    Medium impactLow effort First result in 1 month CTO

    First action

    Collect the existing answers from the FAQ, /data and /compare into one draft.

    How to measure it

    Security review time

    Now
    –
    Target
    -30%
    By
    2 quarters

    Steps

    1. 1

      Gather answers

      From FAQ, /data, /compare and docs.

      CTO 1 day

    2. 2

      Explain teleport

      Approval, Touch ID, what is never sent.

      CTO 1 day

    3. 3

      Explain telemetry

      What, why and how to opt out.

      CTO half a day

    4. 4

      Add the report path

      SOC 2 Type I on request via the form.

      Founding Technical GTM Lead 1 hour

    5. 5

      Publish and link

      Footer, pricing, labs form, sitemap.

      Founding Engineer, Infra & Agent Systems 1 day

    6. 6

      Keep current

      Update with each security-relevant release.

      CTO ongoing

    First result
    1 month
    Cost
    $0
    Owner
    CTO

    Smallest test

    Send the draft to the next prospect's security reviewer.

    Why it works

    Create /trust with: the SOC 2 Type I report on request, how teleport approval and Touch ID work and what is never transferred, what telemetry collects and how to turn it off, where data lives when you bring your own computers, licences, and how to report a vulnerability. Link it from the footer, pricing and the labs form.

    Security reviews stall deals when answers are spread across pages; one page shortens every enterprise evaluation and answers the press concern head-on.

    Risks and how to handle them

    • Overpromising

      State only what is shipped and audited.

    Evidence

    • Your websiteEnvironments Catalog (/data): "Train and evaluate agents on real computer-use tasks."; "4K+ environments and 400K+ tasks" on Linux, Windows and macOS; scoped data packages with trajectories, annotations and human-reviewed golden runs; a five-environment sample pack; "SOC 2 Type I"; no prices; "Request the full catalog" and "Request scope" open an email to a contact address on the older trycua.com domain. The page is in the sitemap but not linked from the homepage, the pricing page or the footer.Source
    • Press coverageRuntimeWire, October 2, 2026: Cua launched Spaces as a free Mac app for agent desktops people can watch and take over; teleport moves a signed-in app session into a Space only after the user approves it, which the article notes gives the Space's controller "a live authenticated foothold".Source
    • Your websiteHomepage FAQ: teleport reads nothing until the user approves and asks for Touch ID on sensitive items; bundles are checked against a SHA-256 hash; telemetry is on by default and can be turned off; Spaces source-available under FSL-1.1-MIT, Cua Driver and Lume MIT.Source
    • Competitor researchE2B calls itself "The AI agent cloud": SOC 2 Type II, a Hobby tier with one-time credits, Pro at $150 a month plus usage, and an Enterprise page covering bring-your-own-cloud deployments.Source
  • 13

    Cua ships almost every week (27 changelog entries through November 2, 2026), but the blog's last post was August 11, and neither the blog nor the changelog offers a newsletter or RSS link. Spaces, the gateways, Hetzner support and the benchmarks library all launched after the last blog post, so the people who starred the repo have no way to hear about them.

    Medium impactLow effort First result in 1 month Founding Technical GTM Lead

    First action

    Add a signup box and RSS link to /changelog.

    How to measure it

    Subscribers

    Now
    –
    Target
    2,000
    By
    90 days

    Steps

    1. 1

      Add signup and RSS

      Blog and changelog.

      Founding Engineer, Infra & Agent Systems 1 day

    2. 2

      Invite existing accounts

      One opt-in email.

      Founding Technical GTM Lead 1 day

    3. 3

      Write issue one

      What shipped since August.

      Founding Technical GTM Lead 2 days

    4. 4

      Publish a launch post

      Spaces on owned hardware.

      CEO 1 week

    5. 5

      Cross-post

      Discord, X, LinkedIn.

      Founding Technical GTM Lead monthly

    6. 6

      Measure

      Opens, clicks and reactivated users.

      Founding Technical GTM Lead monthly

    First result
    1 month
    Cost
    Email tool cost
    Owner
    Founding Technical GTM Lead

    Smallest test

    Send one "what shipped" email to existing accounts.

    Why it works

    Add an email signup and RSS to the blog and changelog. Send a monthly "What shipped" email built from the changelog plus one benchmark and one user story. Publish one blog post a month on the biggest launch (Spaces on owned hardware, the gateways, 448 sandboxes on one Mac), and cross-post to Discord and X.

    A monthly email reactivates dormant users at almost no cost, and turns shipping speed, Cua's strongest habit, into visible momentum.

    Risks and how to handle them

    • No time to write

      Build the email from changelog entries; one hour a month.

    Evidence

    • Your websiteBlog (31 posts): latest post August 11, 2026; deep posts on macOS, Windows and Linux computer-use internals, plus 2025 posts on Hack the North and a Cua hackathon; no newsletter signup or RSS link. The changelog has 27 dated entries, shipping almost weekly through November 2, 2026.Source
    • Your websiteChangelog, November 2, 2026: "compatibility gateways for E2B and Daytona", Hetzner as your own cloud, an elastic pool, "cua share" and a public relay.Source
  • 14

    Cua's 18 benchmark pages are unusually rigorous (dated, with the machine, raw data and the command to reproduce), but all were published on the same day, October 25, so they made one splash instead of eighteen. Hacker News has rewarded this kind of work: the Lume 0.2 Show HN reached 154 points and 44 comments, with commenters asking about agent traffic visibility and GPU passthrough, which Cua then wrote up.

    High impactMedium effort First result in 1 week Founding Technical GTM Lead

    First action

    Draft the Show HN post around one benchmark and the bring-your-own-computers story.

    How to measure it

    Downloads attributed to HN, X and LinkedIn

    Now
    –
    Target
    1,000
    By
    90 days

    Steps

    1. 1

      Order the series

      Most surprising result first.

      Founding Technical GTM Lead 1 day

    2. 2

      Draft the Show HN

      Plain, technical, with the reproduction command.

      CTO 1 day

    3. 3

      Answer comments

      Engineers in the thread all day.

      CTO 1 day

    4. 4

      Weekly threads

      One chart, one command.

      Founding Technical GTM Lead weekly

    5. 5

      Collect runs

      Community results on each page.

      Founding Engineer, Infra & Agent Systems ongoing

    6. 6

      Track

      Tagged downloads by channel.

      Founding Technical GTM Lead weekly

    First result
    1 week
    Cost
    Staff time only
    Owner
    Founding Technical GTM Lead

    Smallest test

    Post one benchmark thread this week and measure downloads.

    Why it works

    Post a Show HN for Spaces on owned hardware with the reproducible cold-start benchmark as the hook. Then publish one benchmark a week as a short thread on X and LinkedIn with a chart and the reproduction command, inviting people to run it on their own Mac and post results. Collect community runs on each benchmark page.

    Reproducible numbers invite developers to participate rather than just read, and a weekly cadence keeps Cua in front of the same audience for a quarter.

    Risks and how to handle them

    • Comparison claims draw criticism

      Lead with the caveats the pages already state.

    Evidence

    • Your websiteBenchmark "Spaces vs E2B cold start" (measured October 25, 2026): create-to-first-command p50 234.6 ms for a Space on a Mac Studio against 299 ms for E2B over the internet, p99 270.1 ms against 877.8 ms, 30 runs each, with raw CSV and JSON and the commands to reproduce it; the page flags that it is not like-for-like. All 18 benchmark pages are dated October 25, 2026.Source
    • Hacker News"Show HN: Lume 0.2 – Build and Run macOS VMs with unattended setup": 154 points and 44 comments; commenters asked about older macOS installers, agent traffic visibility and GPU passthrough (which the August 2026 blog post covers), and praised the unattended setup and MCP integration.Source
  • 15

    Cua Driver is MIT-licensed and already embedded in agent projects from labs and developer platforms shown on the homepage, and the GTM lead's job includes designing deployments around Driver and launching "agent products built on Cua Driver". Yet the /cua-driver page offers only docs, GitHub and a Spaces download: nothing for a company that wants guaranteed fixes, platform coverage or help shipping it.

    High impactMedium effort First result in 1 quarter Founding Technical GTM Lead

    First action

    List the public projects that embed Cua Driver and rank them by activity.

    How to measure it

    Paid Driver support contracts

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Map embedders

      Public projects that import Driver.

      Founding Technical GTM Lead 1 week

    2. 2

      Define the plan

      Queue, response times, test suite.

      CTO 3 days

    3. 3

      Price it

      Annual tiers by platforms covered.

      CEO 2 days

    4. 4

      Add to /cua-driver

      One section and the labs form.

      Founding Engineer, Infra & Agent Systems 1 day

    5. 5

      Contact maintainers

      Ten conversations.

      Founding Technical GTM Lead 1 month

    6. 6

      Deliver

      Run the test suite for the first customers.

      Founding Engineer, Infra & Agent Systems ongoing

    First result
    1 quarter
    Cost
    Staff time; support capacity
    Owner
    Founding Technical GTM Lead

    Smallest test

    Offer the plan to two embedders this month.

    Why it works

    Add a "Shipping Cua Driver in your product?" section to /cua-driver with an annual support plan: a private issue queue, response-time commitments, early access to platform fixes (Windows, Linux desktops), a compatibility test suite run against their app, and optional co-marketing. Find candidates by reviewing which public projects import the Driver, and contact their maintainers through the GTM lead.

    Companies that build on open-source infrastructure pay to reduce the risk of it breaking in production; this converts usage Cua already has into revenue without changing the licence.

    Risks and how to handle them

    • Support load on a tiny team

      Cap the number of contracts and price accordingly.

    Builds on

    Evidence

    • Your websiteCua Driver page: "Agents use your apps while you keep working.": open-source (MIT) driver for native apps on macOS and Linux, Windows in preview, working with popular coding agents and "any MCP client"; one-line install; no star count shown and no offer for companies that ship it inside their own products.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
    • Your websiteHomepage (checked November 10, 2026): headline "Combine all your computers into one supercomputer."; plan limits (Free: 2 computers in use at once and 20 remote hours a month; Pro: 5 and 200; each Teams seat adds 5 and 200); a "Used in projects from" strip of 15 AI labs, developer platforms and agent projects, each linking to a public code file or doc rather than a story; no testimonials; no link to /data, "enterprise", "contact sales", "SOC 2" or a demo.Source
  • 17

    The catalog offers 4K+ environments, 400K+ tasks and scoped data packages, but no prices and only "Request scope" emails. The GTM post asks for "repeatable offerings, pricing" and "a reliable path from evaluation to paid deployment". A buyer cannot tell whether a pilot costs hundreds or hundreds of thousands, so smaller agent startups may never ask.

    High impactMedium effort First result in 1 quarter Founding Technical GTM Lead

    First action

    Write the scope and price of the Evaluation pilot from the last two catalog deals.

    How to measure it

    Paid pilots

    Now
    –
    Target
    4
    By
    2 quarters

    Steps

    1. 1

      Review past requests

      What buyers asked for and paid.

      Founding Technical GTM Lead 2 days

    2. 2

      Define three pilots

      Scope, deliverables, length.

      Founding Technical GTM Lead 1 week

    3. 3

      Set starting prices

      One number per pilot.

      CEO 2 days

    4. 4

      Publish on /data

      With the sample pack and form.

      Founding Engineer, Infra & Agent Systems 1 day

    5. 5

      Write the conversion path

      Pilot to annual agreement terms.

      Founding Technical GTM Lead 1 week

    6. 6

      Review

      Pilot win and conversion rates.

      Founding Technical GTM Lead monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Founding Technical GTM Lead

    Smallest test

    Quote the Evaluation pilot at a fixed price to the next two inbound requests.

    Why it works

    Define three fixed-scope pilots: an Evaluation pilot (a set of environments in one domain with verifiers, run against the buyer's agent), a Data pilot (trajectories and golden runs for a chosen app set), and a Fleet pilot (Spaces on the buyer's hardware with Driver support). Publish "pilots from $X" on /data with what is included and how long it takes, and convert finished pilots into annual agreements.

    Fixed-scope pilots with a visible starting price shorten procurement and give the GTM lead a repeatable motion instead of custom scoping every time.

    Risks and how to handle them

    • A public price anchors large deals low

      Publish a starting price for the pilot only, not the annual agreement.

    Builds on

    Evidence

    • Your websiteEnvironments Catalog (/data): "Train and evaluate agents on real computer-use tasks."; "4K+ environments and 400K+ tasks" on Linux, Windows and macOS; scoped data packages with trajectories, annotations and human-reviewed golden runs; a five-environment sample pack; "SOC 2 Type I"; no prices; "Request the full catalog" and "Request scope" open an email to a contact address on the older trycua.com domain. The page is in the sitemap but not linked from the homepage, the pricing page or the footer.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
  • 22

    Cua's comparison already holds the inputs: per sandbox-hour prices from $0.0666 to $0.1492 on hosted providers against $0 on owned machines, 448 sandboxes on one Mac Studio, and run-time caps of 24 hours on several hosted Pro plans. But a reader has to do the math themselves to see what it means for their own workload.

    Medium impactMedium effort First result in 1 month Founding Technical GTM Lead

    First action

    Write the formula from the existing price table for three typical workloads.

    How to measure it

    Calculator-to-download rate

    Now
    –
    Target
    15%
    By
    90 days

    Steps

    1. 1

      Write the formula

      From the dated price table and benchmarks.

      Founding Technical GTM Lead 2 days

    2. 2

      Pick defaults

      Three typical agent workloads.

      Founding Technical GTM Lead 1 day

    3. 3

      Build it

      On /compare, with shareable results.

      Founding Engineer, Infra & Agent Systems 1 week

    4. 4

      Add the next step

      Download, plan and labs form.

      Founding Engineer, Infra & Agent Systems 1 day

    5. 5

      Promote

      Comparison pages, newsletter, threads.

      Founding Technical GTM Lead ongoing

    6. 6

      Refresh prices

      With each comparison update.

      Founding Technical GTM Lead quarterly

    First result
    1 month
    Cost
    Engineering time
    Owner
    Founding Technical GTM Lead

    Smallest test

    Publish a static table of three workloads first.

    Why it works

    Build a calculator on /compare: inputs are sandboxes running at once, hours a day and run length; outputs are the monthly hosted bill per provider (from the dated price table), the number of owned machines or Hetzner servers needed, and the Cua plan that fits. End with "Run this on your Mac in five minutes" and the labs form for large fleets. Let people share a link to their result.

    Turning a table into the buyer's own number makes the saving concrete and gives champions something to forward internally.

    Risks and how to handle them

    • Ignores the cost of owning hardware

      Include hardware or Hetzner rental cost so the comparison stays fair.

    Builds on

    Evidence

    • Your website"Cua Spaces and hosted sandboxes compared" (checked October 24, 2026): E2B, Daytona, Modal Sandboxes, Vercel and Cloudflare side by side on price per sandbox-hour (Cua "$0 on your own machines" against $0.0666 to $0.1492 for 1 vCPU and 1 GiB), longest run, concurrency (448 sandboxes on one Mac Studio), cold start and licences; Daytona, Vercel and Cloudflare cold starts are still "To be measured". The /compare hub itself links only to /compare/clouds and the benchmarks.Source
    • Your websiteBenchmark "Spaces vs E2B cold start" (measured October 25, 2026): create-to-first-command p50 234.6 ms for a Space on a Mac Studio against 299 ms for E2B over the internet, p99 270.1 ms against 877.8 ms, 30 runs each, with raw CSV and JSON and the commands to reproduce it; the page flags that it is not like-for-like. All 18 benchmark pages are dated October 25, 2026.Source

Big bets 2

Larger plays that can change the trajectory.

  • 18

    Cua-Bench already produces results the field pays attention to: on 25 expert KiCad tasks the best frontier agent clears 6, and Cua published the work jointly with Snorkel AI. A public leaderboard creates demand from every team whose agent is on it, or wants to be, but the page's only commercial hook is a "get in touch" label with nothing behind it.

    High impactHigh effort First result in 1 quarter CTO

    First action

    Offer one private evaluation to a team already on the public leaderboard.

    How to measure it

    Private evaluation revenue

    Now
    –
    Target
    5 paid runs
    By
    2 quarters

    Steps

    1. 1

      Define the product

      Held-out sets, report, trajectories.

      CTO 2 weeks

    2. 2

      Build the runner

      Submit an agent, run, report.

      Founding Engineer, Infra & Agent Systems 1 month

    3. 3

      Price it

      Per run and quarterly.

      CEO 1 week

    4. 4

      Wire the CTA

      Cua-Bench "get in touch" to the form.

      Founding Engineer, Infra & Agent Systems 1 day

    5. 5

      Add a domain each quarter

      Keep the leaderboard news-worthy.

      Research team quarterly

    6. 6

      Sell

      Teams on and near the leaderboard.

      Founding Technical GTM Lead ongoing

    7. 7

      Feed the catalog

      Failures become new tasks.

      Research team ongoing

    First result
    1 quarter
    Cost
    Engineering and compute; task authoring
    Owner
    CTO

    Smallest test

    Run one paid private evaluation by hand.

    Why it works

    Offer private evaluation runs: a lab or agent company submits its agent, Cua runs it on held-out task sets in the buyer's chosen domains (with the canary rules the page already uses), and returns a report with failure analysis and trajectories. Sell per run or by quarterly subscription, add one new public domain each quarter to keep the leaderboard fresh, and feed the failures into the Environments Catalog.

    Teams pay for independent, repeatable evaluation they cannot easily build themselves, and the public leaderboard is the marketing for the private product.

    Risks and how to handle them

    • Perceived conflict of interest

      Keep public and private task sets separate and publish the method.

    Builds on

    Evidence

    • Your websiteCua-Bench: "Make your agents better at computers."; 25 expert-authored KiCad schematic tasks on which the best frontier agent clears 6 of 25; a joint report with Snorkel AI; an "Interested in" prompt labelled "get in touch" with no form or email behind it.Source
    • Your websiteEnvironments Catalog (/data): "Train and evaluate agents on real computer-use tasks."; "4K+ environments and 400K+ tasks" on Linux, Windows and macOS; scoped data packages with trajectories, annotations and human-reviewed golden runs; a five-environment sample pack; "SOC 2 Type I"; no prices; "Request the full catalog" and "Request scope" open an email to a contact address on the older trycua.com domain. The page is in the sitemap but not linked from the homepage, the pricing page or the footer.Source
    • HiringFounding Technical GTM Lead: owns "the full sales cycle" from sourcing to expansion; sells to AI labs, model providers, agent startups and companies deploying computer-use agents; designs deployments around "Cua Driver, computer fleets, evaluation environments, or trajectory data"; turns early projects into "repeatable offerings, pricing, case studies"; values experience converting open-source adoption into revenue; "This is not a role where you inherit a sales playbook. You'll create it." The post still cites "9k+ GitHub stars in just 4 months".Source
  • 19

    Cua deliberately does not host or rent computers: Spaces runs on your Mac, your other computers or your own AWS, Google Cloud or Modal accounts, and the November 2 release added Hetzner as your own cloud and an elastic pool. Cua's own comparison shows the cost per sandbox-hour on owned or cheap rented machines is far below hosted sandboxes. The missing piece for a team without spare Macs is a fast, trusted way to stand up their own machines.

    High impactHigh effort First result in 1 quarter Founding Technical GTM Lead

    First action

    Ship a Hetzner template and measure time from click to first Space.

    How to measure it

    Template-provisioned teams

    Now
    –
    Target
    100
    By
    2 quarters

    Steps

    1. 1

      Build the Hetzner template

      Pool, Spaces install, team join.

      Founding Engineer, Infra & Agent Systems 3 weeks

    2. 2

      Build the AWS template

      Same flow.

      Founding Engineer, Infra & Agent Systems 3 weeks

    3. 3

      Publish costs

      Cost per Space-hour, benchmarked.

      Founding Engineer, Infra & Agent Systems 1 week

    4. 4

      Apply to partner programmes

      Directories and startup credits.

      Founding Technical GTM Lead 1 month

    5. 5

      Co-market

      Joint post or webinar per partner.

      Founding Technical GTM Lead 1 quarter

    6. 6

      Measure

      Template teams and paid conversion.

      CTO monthly

    First result
    1 quarter
    Cost
    Engineering time; partner programme effort
    Owner
    Founding Technical GTM Lead

    Smallest test

    Ship the Hetzner template to ten Discord members and time their setup.

    Why it works

    Build one-click templates (a launch button or a single CLI command) that create a right-sized pool of machines in the customer's own Hetzner or AWS account, pre-installed with Spaces and connected to their team. Approach each provider's startup or partner programme for co-marketing, a listing in their partner directory and credits for new Cua teams. Publish the cost per Space-hour for each template alongside the existing benchmarks.

    It keeps the no-hosting position while removing the main reason a team would pick a hosted sandbox instead: not having machines ready. Cloud partners gain usage, so they have a reason to promote it.

    Risks and how to handle them

    • Support burden for customer clouds

      Support only the templates' default configuration.

    Evidence

    • Your website"Bring the computers. Spaces does the rest.": Cua does not host, rent or resell computers; Spaces runs on your Mac, your other computers or your own AWS, Google Cloud or Modal accounts; the relay is free up to 2 computers in use and 20 remote hours a month; no contact path.Source
    • Your websiteChangelog, November 2, 2026: "compatibility gateways for E2B and Daytona", Hetzner as your own cloud, an elastic pool, "cua share" and a public relay.Source
    • Your website"Cua Spaces and hosted sandboxes compared" (checked October 24, 2026): E2B, Daytona, Modal Sandboxes, Vercel and Cloudflare side by side on price per sandbox-hour (Cua "$0 on your own machines" against $0.0666 to $0.1492 for 1 vCPU and 1 GiB), longest run, concurrency (448 sandboxes on one Mac Studio), cold start and licences; Daytona, Vercel and Cloudflare cold starts are still "To be measured". The /compare hub itself links only to /compare/clouds and the benchmarks.Source

Wildcards 2

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

  • 20

    Sharing is built into Spaces: the Free plan already lets you "Share a Space with named people", and "cua share" shipped on November 2. Every share puts the product in front of a new person, and remote hours (20 a month on Free) are the scarce resource, so they make a natural reward.

    Medium impactMedium effort First result in 1 month CTO

    First action

    Count how many Spaces are shared today and how many invitees sign in.

    How to measure it

    Share-driven sign-ups

    Now
    –
    Target
    15% of new accounts
    By
    2 quarters

    Steps

    1. 1

      Baseline

      Shares and invitee sign-ins.

      Founding Engineer, Infra & Agent Systems 2 days

    2. 2

      Set the reward

      Hours, cap and abuse rules.

      CTO 1 day

    3. 3

      Build it

      Share dialog and welcome screen.

      Founding Engineer, Infra & Agent Systems 2 weeks

    4. 4

      Add a Teams credit

      A month of a seat on conversion.

      CTO 1 week

    5. 5

      Announce

      Changelog, email and Discord.

      Founding Technical GTM Lead 1 day

    6. 6

      Measure

      Invites per sharer and paid conversion.

      CTO monthly

    First result
    1 month
    Cost
    Relay capacity
    Owner
    CTO

    Smallest test

    Show the reward copy in the share dialog to half of users for a month.

    Why it works

    When someone shares a Space and the invitee signs in and opens it, give both 10 extra remote hours that month (capped). Show the reward in the share dialog and in the invitee's welcome screen, and track invites per sharer. For Teams, credit a month of a seat when an invitee's company starts a Teams plan.

    The reward is cheap to give (relay hours) and is tied to the action that already demonstrates the product best: watching an agent work in a shared desktop.

    Risks and how to handle them

    • Self-referral abuse

      Require distinct accounts with a verified sign-in and cap rewards.

    Builds on

    Evidence

    • Your websitePricing page: Free $0; Pro "$15/mo" billed yearly ($180) or $18 monthly; Teams "$25/seat/mo" yearly or $30 monthly, "1 to 100 seats", with shared team computers, roles and device approval; "Share links that expire" and "90-day audit history" marked "Coming soon", SSO and audit export "coming next"; no trial; over a limit "nothing is cut off" and the app shows an Upgrade notice; no enterprise tier, contact-sales option or email, only a link to /support.Source
    • Your websiteChangelog, November 2, 2026: "compatibility gateways for E2B and Daytona", Hetzner as your own cloud, an elastic pool, "cua share" and a public relay.Source
    • Your websiteHomepage (checked November 10, 2026): headline "Combine all your computers into one supercomputer."; plan limits (Free: 2 computers in use at once and 20 remote hours a month; Pro: 5 and 200; each Teams seat adds 5 and 200); a "Used in projects from" strip of 15 AI labs, developer platforms and agent projects, each linking to a public code file or doc rather than a story; no testimonials; no link to /data, "enterprise", "contact sales", "SOC 2" or a demo.Source
  • 21

    Cua has run hackathons before (posts on Hack the North and a Cua hackathon in 2025), publishes research-grade work such as its NeurIPS paper round-up and Cua-Bench, and keeps a research intern role open. Students and researchers who learn computer-use on Cua become the engineers who choose infrastructure at labs and agent startups later.

    Medium impactMedium effort First result in 1 quarter Research team

    First action

    Pick two hackathons in the next quarter and offer a Cua track.

    How to measure it

    Active education accounts

    Now
    –
    Target
    300
    By
    2 quarters

    Steps

    1. 1

      Define eligibility

      University email or event code.

      CTO 1 day

    2. 2

      Build the kit

      Template, task set, Driver example.

      Research team 2 weeks

    3. 3

      Pick events

      Two hackathons and three labs.

      Founding Technical GTM Lead 1 week

    4. 4

      Run a track

      Prize for the best Cua project.

      Research team per event

    5. 5

      Tag accounts

      For later follow-up.

      Founding Engineer, Infra & Agent Systems 1 day

    6. 6

      Recruit

      Invite top builders to the intern pipeline.

      CTO ongoing

    First result
    1 quarter
    Cost
    Prize money and relay hours
    Owner
    Research team

    Smallest test

    Sponsor one hackathon track this quarter.

    Why it works

    Offer free Pro for a semester to verified university labs and hackathon teams, a starter kit (a Space template, a Cua-Bench task set, a Driver example), and a small prize for the best project built on Cua. Tag these accounts so the GTM lead can follow graduates who move to companies, and use the programme as the pipeline for the research intern role.

    Education programmes are cheap for software with no hosting cost and seed long-term adoption in exactly the community Cua sells to.

    Risks and how to handle them

    • No near-term revenue

      Keep costs small and measure hiring and later conversions.

    Evidence

    • Your websiteBlog (31 posts): latest post August 11, 2026; deep posts on macOS, Windows and Linux computer-use internals, plus 2025 posts on Hack the North and a Cua hackathon; no newsletter signup or RSS link. The changelog has 27 dated entries, shipping almost weekly through November 2, 2026.Source
    • HiringCua's Y Combinator jobs page (cua.ai/careers redirects there): batch P25, team size 3, tagline "Give every agent a cloud desktop."; three open roles: Founding Technical GTM Lead ($150K-$200K, 0.50%-1.50% equity), Founding Engineer, Infra & Agent Systems ($100K-$150K) and "Research Intern (Summer 2026)", still listed in November 2026.Source
    • Your websiteCua-Bench: "Make your agents better at computers."; 25 expert-authored KiCad schematic tasks on which the best frontier agent clears 6 of 25; a joint report with Snorkel AI; an "Interested in" prompt labelled "get in touch" with no form or email behind it.Source

03

How Programmable can help

Where an idea in this playbook is something Programmable builds or runs, it's listed here with what we'd do. Everything else is yours to run.

ProgrammableInfo

Turn the research into a list you can act on

These ideas start with finding the right companies. Info finds them and the right person at each, researches every one and drafts the outreach, then asks before it sends.

  • Build this list in Info for idea #15 (opens programmable.info in a new tab)

New accounts: 100,000 free credits with code BETA100K

Find the companies (opens programmable.info in a new tab)
ProgrammableTalk

Fill the events in this plan

These ideas are events. Talk gets registrants for them and tells you who's coming and what their company does before the day.

Plan an event (opens programmable.talk in a new tab)
More example playbooks and reviews
ProgrammableDigital

Show Cua’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.

  • GhostEye growth strategyAn automated social engineering and breach-and-attack-simulation platform: voice, SMS, email and chat impersonation tests (help-desk…22 growth ideas
  • Albeado growth strategyA Silicon Valley causal AI company whose PRISM platform combines knowledge search with citations, automated reports (IC memos, DD reports…21 growth ideas
  • Lightspeed Systems growth strategyCloud software for K-12 schools: web filtering, student safety monitoring, classroom management, digital hall passes, device management…21 growth ideas
  • Office Hours growth strategyAn on-demand expert network: clients book 30 to 60 minute calls and surveys with vetted experts, and Expert Data Labs supplies…22 growth ideas
  • Zeely growth strategyAn agentic AI platform for making, launching and scaling paid ads: UGC-style video, static and image ads, AI avatars, bulk Meta launches…23 growth ideas
  • Houghton Horns growth strategyA family-owned brass specialist in Keller, Texas: new and pre-owned horns, trombones and trumpets, mouthpieces, cases and sheet music, a…21 growth ideas
  • Beyond Border growth strategyA Singapore-founded U.S. immigration services company that prepares O-1, EB-1A, EB-2 NIW, L-1 and E-2 cases and pairs clients with vetted…20 growth ideas
  • Molagri growth strategyEngineers species-specific biopesticides from insect virus-like particles, built so the targeting element or the payload can be swapped…23 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. Cua did not commission, review or endorse it, and Programmable is not affiliated with Cua. 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 Cua runs its business.

Something wrong, or do you represent Cua? 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 306 ideas

    Open the sales door: link the Environments Catalog and a talk-to-us plan, route contact, fix the sitemap, refresh recruiting copy and align the README pricing; quick image fix.

  2. Days 31 to 609 ideas

    Proof and comparison: proof stories, rival and switching pages, a reason to create an account, upgrade nudges, a trust page, a newsletter, a benchmark series and a cost calculator.

  3. Days 61 to 907 ideas

    Package revenue: an enterprise plan, a Driver commercial offer, catalog pilot packages, then the big bets and wildcards.

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 many agent desktops on hardware they already own, with nothing rented. The most cited problem: hosted sandbox bills and run-time caps.

  • Agents that use native apps without taking over the person's cursor
  • Agents stopping at permission dialogs and sign-ins
  • In their words: “computer-use”

Market

Competitive landscape

4 competitors compared: E2B, Daytona, Modal Sandboxes and Vercel and Cloudflare sandboxes.

  • Gap to exploit: Full macOS, Windows and Linux desktops for agents on hardware the team owns
  • E2B: Hosted and metered per sandbox-hour; Linux only
  • Daytona: No full desktops for computer use

Hiring signals

Cua is a three-person YC company hiring its first go-to-market lead and its first infrastructure engineer.

  • Founding Technical GTM Lead: Sales to AI labs and agent companies; offerings, pricing and case studies. Ideas 1, 6, 11, 15, 17.
  • Founding Engineer, Infra & Agent Systems: Multi-cloud orchestration and developer APIs. Ideas 8, 19.
  • Research Intern (Summer 2026): Research pipeline, still listed after the summer. Ideas 4, 21.

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 firstNew accounts get 100,000 free enrichment credits with code BETA100K (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)
  • ProgrammableMediaAn AI note taker, interpreter and webinar host for your calls (opens programmable.media 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