e2e (by TesterArmy, Inc.) growth playbook

Publish e2e.com/cloud with one paragraph on what the cloud will do (run e2e suites on managed browsers, simulators and devices, in parallel, for pull requests) and a short waitlist form. The form asks for work email, team size, platforms (web, iOS, Android), current CI, current tool (Playwright, Cypress, Detox, Maestro, TesterArmy, none), the number of end-to-end tests, and whether the team runs e2e today. Link it from the seed post, the enterprise page, the README, the CI docs and the end of `e2e init` output as a single optional line. Review answers weekly, score teams by size and fit, and invite the best to design-partner calls.

A waitlist turns announced intent into a list the company owns, at the moment interest peaks after the raise. The answers tell the team which runners, devices and CI vendors to build for first. The list also gives the new Chief of Staff the pipeline numbers for the revenue model and investor reporting the role owns.

[ ] Work email
[ ] Team size: 1-9 / 10-49 / 50-199 / 200+
[ ] Platforms: web / iOS / Android
[ ] CI: GitHub Actions / GitLab / EAS Workflows / Bitrise / Codemagic / other
[ ] Current tool: Playwright / Cypress / Selenium / Detox / Maestro / TesterArmy / none
[ ] Number of end-to-end tests today
[ ] Running e2e already? yes / no
[ ] Open to a 20-minute design-partner call? yes / no

• Developers see a commercial push in an open-source CLI

  Keep it to one line in init, never in run or CI output, and say clearly that the framework stays Apache-2.0.

• A long wait cools interest

  Send a short monthly build update and open early access in waves.

• Your websiteSeed announcement (November 3, 2026): "$8 million" from Credo Ventures, KAYA, Point Nine, Baobab Ventures and Script Capital; in the first month after the October 1 launch e2e "passed 1.5M downloads on npm and 12k stars on GitHub, and welcomed 60 external contributors" and "now gets around 100,000 downloads a day"; "we're launching e2e cloud, so teams can run e2e on the best infrastructure" next year; "More tools will follow for every software verification step"; the framework stays Apache-2.0.Source
• Your websiteSitemap (160 URLs, checked November 9, 2026): 58 examples, 74 docs pages, 6 blog posts, 5 sessions pages, 4 comparison pages (Playwright, Maestro, Detox, Selenium) and 7 migration guides (Playwright, Cypress, Selenium, Appium, Detox, Maestro, TesterArmy), plus enterprise, contact, team and three careers pages; no pricing, cloud, waitlist, security or customers page.Source
• Your websiteContact page: full name, work email, optional company website and country, a topic menu (Enterprise, Partnerships, Open Source, Other) and an optional message; the submit button says "book a call" but the page shows no calendar or scheduling step and no response time.Source
• Your websiteHomepage: "The open source AI testing framework. Customizable TypeScript testing framework for web, mobile apps and more."; starts with `npx e2e@latest init`; sections compare e2e with Playwright, Cypress and Selenium, Maestro and Detox, "Playwright MCP and agent-only AI testing" and TesterArmy; no email signup or waitlist.Source
• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source

Write /compare/cypress in the same format as the other four: what Cypress is, the same test in both, differences, a feature table that marks where Cypress leads (component testing, its interactive runner, its cloud dashboard), and when to keep Cypress. Lead with what Cypress cannot do: native iOS and Android, agent steps and coding-agent workflows. End with the migration guide and a "book a call" for larger suites. Repeat for Appium next.

Comparison searches come from people already choosing, and the page reuses a format, a guide and a fact-checking process that already exist.

• Claims go stale as Cypress ships

  Date the page and keep the GitHub corrections link.

• Your websiteComparison hub: four pages (Playwright, Selenium, Maestro, Detox); Cypress has no comparison page, only a pointer to the Cypress migration guide; Appium and WebdriverIO are folded into other guides; agent-first QA products are not compared; facts are "from public documentation and release notes", rows where a rival leads are marked, and corrections are taken on GitHub.Source
• Your websiteHomepage: "The open source AI testing framework. Customizable TypeScript testing framework for web, mobile apps and more."; starts with `npx e2e@latest init`; sections compare e2e with Playwright, Cypress and Selenium, Maestro and Detox, "Playwright MCP and agent-only AI testing" and TesterArmy; no email signup or waitlist.Source
• Your websiteSitemap (160 URLs, checked November 9, 2026): 58 examples, 74 docs pages, 6 blog posts, 5 sessions pages, 4 comparison pages (Playwright, Maestro, Detox, Selenium) and 7 migration guides (Playwright, Cypress, Selenium, Appium, Detox, Maestro, TesterArmy), plus enterprise, contact, team and three careers pages; no pricing, cloud, waitlist, security or customers page.Source

Add a proof section to /enterprise: the team's track record (named TesterArmy users and results, with their permission refreshed for e2e), the Expo integration with a link to Expo's guide, and the usage numbers already on the page. Add one short rollout story as soon as the first enterprise migration finishes. Keep example-app logos separate and labelled as integrations, as the examples page already does.

Buyers of a paid rollout want evidence that the team has done it before. The proof exists but sits in a funding post and a partner's docs.

• TesterArmy results are about the old hosted product

  Label them as the team's track record and replace them with e2e stories as they arrive.

• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source
• Your websitePre-seed post (September 15, 2026): TesterArmy raised "$1.2M" (backers include Y Combinator, Script Capital, Eight Capital and angels from Vercel, Expo and Resend); teams at Resend, bolt.new and Nando's used TesterArmy daily; Novu "cut flaky tests in half"; Juno reported "shipping PRs 10× faster" and two days of manual testing saved a week.Source
• Expo docsExpo's docs list e2e next to TesterArmy in the Testing section, with a sample EAS workflow for an iOS Simulator on each pull request; they recommend an API key over a subscription login on CI; an `@e2e-dev/eas` integration with Expo Simulators is a limited-access preview.Source
• Your websiteExamples: 58 cards, 10 starter apps (Next.js, Vite, Astro, Angular, Expo, SwiftUI, Compose, Flutter, .NET MAUI, Kotlin Multiplatform) and named product examples such as Clerk, Stripe, Paddle, Shopify, RevenueCat, Webiny, Quivr and Kernel; no "used by" section, and logos are "used only to identify the products each example works with"; `npx e2e init --example <name>`.Source

After submit, show a calendar for Enterprise and Partnerships topics, and say how fast the team answers the rest. Add three optional fields: platforms, current tool and number of end-to-end tests. Route large suites straight to a founder's calendar and send everyone else the quickstart and Discord.

Instant booking keeps a buyer's momentum, and three fields let a small team spend its call time on the largest rollouts.

• Founders' calendars fill with small requests

  Show the calendar only above a suite-size threshold.

• Your websiteContact page: full name, work email, optional company website and country, a topic menu (Enterprise, Partnerships, Open Source, Other) and an optional message; the submit button says "book a call" but the page shows no calendar or scheduling step and no response time.Source
• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source
• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source

Email every active TesterArmy account with a dated plan: a free conversion of their saved tests into an e2e repo branch, run by the team with the guide's agent prompt, plus a 30-minute CI setup call. Track each account's status, and fix the most-used missing features first. Offer paid rollout terms to the largest accounts.

Converting the tests for the customer removes the work that causes churn, and each finished migration is a ready case study for the enterprise page.

• A missing feature blocks a large account

  Prioritize gaps by account size and offer a custom build.

• TesterArmyTesterArmy is "winding down": new sign-ups are closed, plans run until January 15, 2027 with nothing billed after, a data export is promised "shortly", and tests must be rewritten as e2e tests; the team is "doubling down on open source".Source
• Your websiteTesterArmy migration guide: a coding agent reads saved tests through the TesterArmy MCP server and rewrites them as e2e tests; features marked "Not yet" or "No equivalent" include inbox credentials, microphone steps, accessibility audits and browser region settings.Source
• Your websitePre-seed post (September 15, 2026): TesterArmy raised "$1.2M" (backers include Y Combinator, Script Capital, Eight Capital and angels from Vercel, Expo and Resend); teams at Resend, bolt.new and Nando's used TesterArmy daily; Novu "cut flaky tests in half"; Juno reported "shipping PRs 10× faster" and two days of manual testing saved a week.Source

Change the release workflow so the main e2e package is always published last and marked latest, and sub-package releases are not marked latest. Add one human-written highlight line to the top of each e2e release, linking to the blog post.

The release label is often the first version number an evaluator sees on a 13.5k-star repo. Showing the real version and its highlights makes the project look as active as it is.

• Release tooling overrides the flag

  Add a check that fails the job if the wrong release is latest.

• GitHubReleases page (checked November 9, 2026): [email protected] shipped on November 5 alongside seven @e2e-dev package releases, and GitHub's "Latest" label sits on @e2e-dev/[email protected], not the main e2e package.Source
• GitHubRepository (checked November 9, 2026): 13.5k stars, 709 forks, 37 open issues, 44 open pull requests, 1,131 commits on main, Apache-2.0, described as "Next generation e2e testing framework for web and mobile apps."Source

Publish e2e's MCP server and skill in the main MCP registries and skill directories, with a clear one-line install, a 30-second demo clip and a link to the coding-agents docs. Track installs by directory with a source tag in the install docs.

Listings put e2e in front of developers at the moment their agent needs a way to test, which is the job e2e is designed for.

• Listings go stale

  Add listing updates to the release checklist.

• HiringOpen Source Developer (worldwide): "design and ship framework features and fixes, from the first issue to the release notes"; review and merge contributor pull requests and answer GitHub and Discord questions; improve e2e inside coding agents (the MCP server, the skill `e2e init` installs, failure traces); write the docs, examples and migration guides for each change.Source
• Your websitee2e 0.23 (November 5, 2026): agent steps on a Claude Code subscription, BrowserStack real devices via `@e2e-dev/browserstack`, an `e2e doctor` preflight check, a `run_tests` MCP tool for coding agents, an engine protocol over JSON-RPC, and startup 25-44% faster.Source

Write three fair pages in the existing format: e2e vs Momentic, e2e vs QA Wolf and e2e vs Maestro Cloud. Be honest about where they lead (hosted runs, managed service, dashboards, compliance), and end with when to choose each. Link the cloud waitlist for teams that want hosted runs.

Teams choosing an AI testing vendor search these names, and an open-source option with no data leaving their infrastructure is a clear position against all three.

• Claims about rivals go out of date

  Date each page and keep the corrections process.

• #2 Add the missing e2e vs Cypress page: the homepage and a migration guide already make the case

• Your websiteComparison hub: four pages (Playwright, Selenium, Maestro, Detox); Cypress has no comparison page, only a pointer to the Cypress migration guide; Appium and WebdriverIO are folded into other guides; agent-first QA products are not compared; facts are "from public documentation and release notes", rows where a rival leads are marked, and corrections are taken on GitHub.Source
• Competitor researchRivals' own homepages: Maestro, a free open-source CLI plus Maestro Studio and paid Maestro Cloud for parallel runs, used by "thousands of teams"; Momentic, an AI QA agent, "Free to start, no credit card", SOC 2 Type 2, enterprise SSO, RBAC and SLA, logos such as Notion, Webflow and Retool; QA Wolf, an "AI Test Factory" with run infrastructure and "Coverage as a Service", naming DoorDash and Figma.See the research Source
• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source

Publish a "what does a run cost" docs page and a small calculator. Show model calls and tokens for first runs, replays and UI-change re-runs on three example apps, priced at a few common providers, plus a local-model option and `--strict-cache` for zero model calls in CI. Add an aggregate figure from telemetry, such as the median share of steps served from cache, once it is stable.

Turning an unknown cost into a small, predictable number removes the main reason engineering leads hesitate to put agent steps in every pull request.

• Provider prices change

  Show calls and tokens first, and date the prices.

• Your websitee2e vs Playwright page: Playwright leads on five languages, component tests, its trace viewer, UI mode, codegen and VS Code extension ("e2e has no graphical viewer"); e2e leads on native mobile, agent steps, exploratory runs and coding-agent workflows; e2e costs "CI minutes plus model calls" when a goal has no recording, and `--strict-cache` stops model calls in CI.Source
• Your websitee2e and Expo (November 9, 2026): tests run on Expo Simulators, one session per worker, about three minutes to start; an EAS Workflows file runs the suite on every pull request; the example's first run took 31.6 seconds and 11 model calls, the cached replay 9.7 seconds and 2 model calls.Source
• Your websiteTelemetry docs: anonymous telemetry is on by default and records a random install id, a hash of the repository's root commit, whether a run is in CI (counted under the CI vendor), test and step counts, cache behaviour and model-call and token totals, never test content; events go to a PostHog project "read in aggregate"; opt-out via `e2e telemetry disable` or DO_NOT_TRACK=1; `E2E_TELEMETRY_FLEET` for platforms.Source
• Your websiteLaunch post (October 1, 2026): `npx e2e init` gets most projects to a first passing test in about a minute; a cached step "costs no tokens"; any model provider, local models, gateways or existing subscriptions with "no added markup"; pre-1.0, so parts of the core API may change; desktop not yet supported.Source

Define activation as a repository with passing e2e runs in CI on at least five different days. Build an aggregate funnel from install to first local pass, first CI run and repeated CI runs, split by engine, CI vendor and version. Find the step where most repositories stop, fix it in the docs and CLI (for example, `e2e doctor` hints for the CI case), and report the activated-repo count weekly.

Teams running e2e on every pull request are the ones who will pay for cloud runs and rollouts, so this measure tracks future revenue better than stars or downloads.

• Opt-outs make counts low

  Track the trend, not absolute numbers.

• Your websiteTelemetry docs: anonymous telemetry is on by default and records a random install id, a hash of the repository's root commit, whether a run is in CI (counted under the CI vendor), test and step counts, cache behaviour and model-call and token totals, never test content; events go to a PostHog project "read in aggregate"; opt-out via `e2e telemetry disable` or DO_NOT_TRACK=1; `E2E_TELEMETRY_FLEET` for platforms.Source
• HiringSoftware Engineer (worldwide): builds the agent behind agent.act, agent.assert, agent.waitFor and agent.extract and measures changes before shipping; makes runs faster and cheaper through caching and replay; "help engineering teams adopt e2e, from their first suite through running it in CI on every pull request"; bonus for building LLM evals.Source
• Your websiteSeed announcement (November 3, 2026): "$8 million" from Credo Ventures, KAYA, Point Nine, Baobab Ventures and Script Capital; in the first month after the October 1 launch e2e "passed 1.5M downloads on npm and 12k stars on GitHub, and welcomed 60 external contributors" and "now gets around 100,000 downloads a day"; "we're launching e2e cloud, so teams can run e2e on the best infrastructure" next year; "More tools will follow for every software verification step"; the framework stays Apache-2.0.Source

Turn the five phases into two or three fixed-scope packages with starting prices. For example: a two-week pilot (one app, 20 critical flows in CI), a migration package (an existing Playwright, Cypress, Detox or Maestro suite up to a set size) and an ongoing support retainer. Publish "from" prices and what is included. Keep custom builds on request.

Fixed packages make budget approval easy for engineering managers and make deals faster and more predictable for a small team.

• Under-pricing complex migrations

  Cap suite size per package and quote above it.

• #4 Make the "book a call" button actually book a call, and ask about the suite

• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source
• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source
• Your websiteTeam page: six people (co-founder and CEO, co-founder and CTO, growth, founding developer relations, design, engineering); "60 external contributors".Source

Run one live session a month on a set theme (CI on every PR, migrating from Detox or Playwright, coding agents writing tests, mobile on real devices), co-hosted with a partner where possible. Add a registration form (email, company, platform) and send the recording to every registrant. Cut each session into short clips for X and the docs.

Live sessions with partners reach their audiences, and registrations become a list of named teams for rollouts and the cloud.

• Low attendance

  Co-host with a partner who promotes it to their audience.

• Your websiteSessions: one live 38-minute intro (October 22, 2026) and four recordings, including a 60-minute "Agentic CI for Expo apps" with Expo and an 8-minute "Introduction to Verification Engineering" by an Expo engineer; no registration or reminder signup.Source
• Your websiteTeam page: six people (co-founder and CEO, co-founder and CTO, growth, founding developer relations, design, engineering); "60 external contributors".Source

Publish /partners listing each integration with what it adds (real devices, hosted browsers, simulators, sandboxes). Give each partner a co-marketing kit: a joint example, a docs page on their side, a session slot and a launch post. Track installs of each @e2e-dev package as the partner's measure.

Each partner has its own developer audience, and a working integration gives them a reason to promote e2e that costs them nothing.

• Partners compete with e2e cloud later

  Position partners as infrastructure that e2e cloud can also run on.

• Your websiteIntegrations and CI docs: Kernel, Lightpanda, EAS, smol machines, Limrun, TestMu, BrowserStack and E2B backends, and CI guides for GitHub Actions, GitLab, EAS Workflows, Bitrise, Codemagic and Norrix.Source
• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source
• Your websiteContact page: full name, work email, optional company website and country, a topic menu (Enterprise, Partnerships, Open Source, Other) and an optional message; the submit button says "book a call" but the page shows no calendar or scheduling step and no response time.Source

Make a README badge ("tested with e2e", linking to e2e.com), and offer it in `e2e init` output and the docs. Find public repositories that depend on e2e, ask their maintainers to add the badge and opt into a /used-by page, and feature one project a week on X.

Badges turn each adopting repository into a link and a recommendation seen by its own contributors and users, and a used-by wall gives the proof enterprise visitors look for.

• Listing projects without consent

  Only show projects whose maintainers opt in.

• Your websiteExamples: 58 cards, 10 starter apps (Next.js, Vite, Astro, Angular, Expo, SwiftUI, Compose, Flutter, .NET MAUI, Kotlin Multiplatform) and named product examples such as Clerk, Stripe, Paddle, Shopify, RevenueCat, Webiny, Quivr and Kernel; no "used by" section, and logos are "used only to identify the products each example works with"; `npx e2e init --example <name>`.Source
• GitHubRepository (checked November 9, 2026): 13.5k stars, 709 forks, 37 open issues, 44 open pull requests, 1,131 commits on main, Apache-2.0, described as "Next generation e2e testing framework for web and mobile apps."Source

Publish a review promise (first response within two working days), label pull requests by area, and run a weekly triage. Add "good first example" issues for frameworks and services not yet in the 58 examples, so contributors extend the catalogue. Thank merged contributors in each release, as the release notes already do.

A fast first review turns one-off contributors into regulars and keeps the project's reputation strong as usage grows.

• Review load slows core work

  Rotate triage and close stale PRs kindly.

• GitHubRepository (checked November 9, 2026): 13.5k stars, 709 forks, 37 open issues, 44 open pull requests, 1,131 commits on main, Apache-2.0, described as "Next generation e2e testing framework for web and mobile apps."Source
• HiringOpen Source Developer (worldwide): "design and ship framework features and fixes, from the first issue to the release notes"; review and merge contributor pull requests and answer GitHub and Discord questions; improve e2e inside coding agents (the MCP server, the skill `e2e init` installs, failure traces); write the docs, examples and migration guides for each change.Source
• Your websiteExamples: 58 cards, 10 starter apps (Next.js, Vite, Astro, Angular, Expo, SwiftUI, Compose, Flutter, .NET MAUI, Kotlin Multiplatform) and named product examples such as Clerk, Stripe, Paddle, Shopify, RevenueCat, Webiny, Quivr and Kernel; no "used by" section, and logos are "used only to identify the products each example works with"; `npx e2e init --example <name>`.Source

Publish /security, drawing on the existing security and telemetry docs: the data-flow diagram (what stays on the customer's machines, what goes to their chosen model provider, what anonymous telemetry sends), how to disable telemetry in CI, the vulnerability reporting process, the licence, and the compliance roadmap for the coming cloud. Add a downloadable answers pack for common questionnaires.

Security review is where open-source tools stall inside larger companies. Answering it in public shortens deals and is a real advantage over hosted rivals.

• Promising certifications too early

  State only what is true today and date the roadmap.

• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source
• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source
• Your websiteTelemetry docs: anonymous telemetry is on by default and records a random install id, a hash of the repository's root commit, whether a run is in CI (counted under the CI vendor), test and step counts, cache behaviour and model-call and token totals, never test content; events go to a PostHog project "read in aggregate"; opt-out via `e2e telemetry disable` or DO_NOT_TRACK=1; `E2E_TELEMETRY_FLEET` for platforms.Source
• Competitor researchRivals' own homepages: Maestro, a free open-source CLI plus Maestro Studio and paid Maestro Cloud for parallel runs, used by "thousands of teams"; Momentic, an AI QA agent, "Free to start, no credit card", SOC 2 Type 2, enterprise SSO, RBAC and SLA, logos such as Notion, Webflow and Retool; QA Wolf, an "AI Test Factory" with run infrastructure and "Coverage as a Service", naming DoorDash and Figma.See the research Source

Propose to Expo an opt-in e2e test step in its project templates and a ready EAS Workflows template that runs e2e on every pull request against Expo Simulators. Build and maintain the templates, add a joint landing page, and agree shared metrics (projects using the template, PR runs).

Defaults decide which test tool a new project uses, and Expo's audience is the mobile-first, agent-heavy developer e2e is built for.

• Expo prefers to stay neutral

  Offer an opt-in choice among tools, with e2e first by merit.

• #13 Launch the partner program with the backends already integrated

• Expo docsExpo's docs list e2e next to TesterArmy in the Testing section, with a sample EAS workflow for an iOS Simulator on each pull request; they recommend an API key over a subscription login on CI; an `@e2e-dev/eas` integration with Expo Simulators is a limited-access preview.Source
• Your websitee2e and Expo (November 9, 2026): tests run on Expo Simulators, one session per worker, about three minutes to start; an EAS Workflows file runs the suite on every pull request; the example's first run took 31.6 seconds and 11 model calls, the cached replay 9.7 seconds and 2 model calls.Source
• Your websiteSessions: one live 38-minute intro (October 22, 2026) and four recordings, including a 60-minute "Agentic CI for Expo apps" with Expo and an 8-minute "Introduction to Verification Engineering" by an Expo engineer; no registration or reminder signup.Source
• Your websitePre-seed post (September 15, 2026): TesterArmy raised "$1.2M" (backers include Y Combinator, Script Capital, Eight Capital and angels from Vercel, Expo and Resend); teams at Resend, bolt.new and Nando's used TesterArmy daily; Novu "cut flaky tests in half"; Juno reported "shipping PRs 10× faster" and two days of manual testing saved a week.Source

Open-source the eval suite the team already needs: a fixed set of web and mobile apps and tasks, scored for success rate, steps, time, model calls and cost per provider and model, with first-run and replay figures. Publish a dated leaderboard on e2e.com and accept community submissions. Use it to set the default model recommendation in the docs.

An open benchmark becomes the reference other people cite, and it answers the practical question every new user asks: which model should I use?

• Seen as self-serving

  Open the harness and data, and let others reproduce results.

• HiringSoftware Engineer (worldwide): builds the agent behind agent.act, agent.assert, agent.waitFor and agent.extract and measures changes before shipping; makes runs faster and cheaper through caching and replay; "help engineering teams adopt e2e, from their first suite through running it in CI on every pull request"; bonus for building LLM evals.Source
• Your websiteLaunch post (October 1, 2026): `npx e2e init` gets most projects to a first passing test in about a minute; a cached step "costs no tokens"; any model provider, local models, gateways or existing subscriptions with "no added markup"; pre-1.0, so parts of the core API may change; desktop not yet supported.Source
• Competitor researchRivals' own homepages: Maestro, a free open-source CLI plus Maestro Studio and paid Maestro Cloud for parallel runs, used by "thousands of teams"; Momentic, an AI QA agent, "Free to start, no credit card", SOC 2 Type 2, enterprise SSO, RBAC and SLA, logos such as Notion, Webflow and Retool; QA Wolf, an "AI Test Factory" with run infrastructure and "Coverage as a Service", naming DoorDash and Figma.See the research Source

Pick ten design partners from the waitlist across web-only, mobile-heavy and mixed teams. Run their suites on an early cloud for a quarter at no or low cost in exchange for weekly feedback and usage data. Test two pricing units (parallel workers per month versus test minutes) with the partners. Launch with a free tier for open-source projects and published prices.

Pricing built from real suites and real willingness to pay avoids the most common open-source mistake: a cloud that costs too much for adopters or too little to fund the company.

• Partners expect free forever

  Agree launch pricing and a discount in the contract.

• #1 Open an e2e cloud waitlist: the seed post promises the cloud, but no page on e2e.com lets a team ask for it
• #10 Make "first green run in CI" the activation metric, from telemetry already collected

• Your websiteSeed announcement (November 3, 2026): "$8 million" from Credo Ventures, KAYA, Point Nine, Baobab Ventures and Script Capital; in the first month after the October 1 launch e2e "passed 1.5M downloads on npm and 12k stars on GitHub, and welcomed 60 external contributors" and "now gets around 100,000 downloads a day"; "we're launching e2e cloud, so teams can run e2e on the best infrastructure" next year; "More tools will follow for every software verification step"; the framework stays Apache-2.0.Source
• HiringChief of Staff (worldwide, open now): owns "security reviews, questionnaires, compliance, data protection, legal terms", "revenue and the P&L", the financial model and investor reporting, and special projects "such as a partner program or a sales change"; bonus for SOC 2 or ISO 27001 and billing-stack experience. Careers banner: "e2e raised $8m to build the software verification layer for the agentic coding era."Source
• Competitor researchRivals' own homepages: Maestro, a free open-source CLI plus Maestro Studio and paid Maestro Cloud for parallel runs, used by "thousands of teams"; Momentic, an AI QA agent, "Free to start, no credit card", SOC 2 Type 2, enterprise SSO, RBAC and SLA, logos such as Notion, Webflow and Retool; QA Wolf, an "AI Test Factory" with run infrastructure and "Coverage as a Service", naming DoorDash and Figma.See the research Source
• Your websiteTelemetry docs: anonymous telemetry is on by default and records a random install id, a hash of the repository's root commit, whether a run is in CI (counted under the CI vendor), test and step counts, cache behaviour and model-call and token totals, never test content; events go to a PostHog project "read in aggregate"; opt-out via `e2e telemetry disable` or DO_NOT_TRACK=1; `E2E_TELEMETRY_FLEET` for platforms.Source

Build a viewer that opens an e2e report locally (steps, screenshots, video, agent reasoning and cache hits), with an optional one-click upload that gives a shareable link for pull-request comments. Keep the local viewer open source. Make hosted links free with limits, and make them the first feature of e2e cloud, so teams start using the cloud before they run tests there.

It closes the most visible gap against Playwright, and shareable run links put e2e in front of every reviewer on the team, not only the engineer who installed it.

• Uploading reports conflicts with "we never receive your data"

  Keep upload opt-in, say so clearly, and offer self-hosting.

• #19 Price e2e cloud from real usage, with ten design partners from the waitlist

• Your websitee2e vs Playwright page: Playwright leads on five languages, component tests, its trace viewer, UI mode, codegen and VS Code extension ("e2e has no graphical viewer"); e2e leads on native mobile, agent steps, exploratory runs and coding-agent workflows; e2e costs "CI minutes plus model calls" when a goal has no recording, and `--strict-cache` stops model calls in CI.Source
• Your websiteSeed announcement (November 3, 2026): "$8 million" from Credo Ventures, KAYA, Point Nine, Baobab Ventures and Script Capital; in the first month after the October 1 launch e2e "passed 1.5M downloads on npm and 12k stars on GitHub, and welcomed 60 external contributors" and "now gets around 100,000 downloads a day"; "we're launching e2e cloud, so teams can run e2e on the best infrastructure" next year; "More tools will follow for every software verification step"; the framework stays Apache-2.0.Source

Set a 1.0 date tied to the cloud launch. Freeze the core test, agent, expect and config APIs, publish a semantic-versioning and deprecation policy (for example, deprecations live for two minor versions and come with codemods), and run a long-term-support line for enterprise rollouts. Announce 1.0 with the first rollout stories.

A stability promise removes a common objection in larger teams and gives the rollout business a firm base.

• Freezing slows innovation

  Keep experimental APIs behind clear labels.

• Your websiteLaunch post (October 1, 2026): `npx e2e init` gets most projects to a first passing test in about a minute; a cached step "costs no tokens"; any model provider, local models, gateways or existing subscriptions with "no added markup"; pre-1.0, so parts of the core API may change; desktop not yet supported.Source
• GitHubReleases page (checked November 9, 2026): [email protected] shipped on November 5 alongside seven @e2e-dev package releases, and GitHub's "Latest" label sits on @e2e-dev/[email protected], not the main e2e package.Source
• Your websiteEnterprise page: "e2e is Apache-2.0 and stays free. You pay only when our engineers help you roll it out or build something custom"; a five-phase rollout (scope, build, migrate, roll out, hand over); "We never receive or process your data"; stats "2.4M npm downloads", "14k GitHub stars", "60 external contributors"; no customer logos or prices, testimonials are posts on X; every call to action is "book a call" to /contact.Source

Add an `expect(screen).toBeAccessible()` style assertion that runs a standard accessibility rule set on web pages, plus an agent check on mobile screens. Include a report section listing issues per screen and step. Write a docs page and an example, and add it as a row on every comparison page.

Accessibility is a requirement many teams must meet. Running it inside tests they already run gives them a new reason to adopt e2e and closes a migration gap at the same time.

• Agent checks on mobile give false positives

  Label mobile checks as advisory until measured.

• #5 Move every TesterArmy account to e2e with a done-for-you conversion before January 15

• Your websiteTesterArmy migration guide: a coding agent reads saved tests through the TesterArmy MCP server and rewrites them as e2e tests; features marked "Not yet" or "No equivalent" include inbox credentials, microphone steps, accessibility audits and browser region settings.Source
• Your websiteComparison hub: four pages (Playwright, Selenium, Maestro, Detox); Cypress has no comparison page, only a pointer to the Cypress migration guide; Appium and WebdriverIO are folded into other guides; agent-first QA products are not compared; facts are "from public documentation and release notes", rows where a rival leads are marked, and corrections are taken on GitHub.Source