Short answer. If you saved the generated Playwright code, running it yourself is the fastest way back to a working suite, and testRigor or Rainforest QA are worth comparing if you want another AI platform to write new tests instead. Octomind confirmed the shutdown in its own archived closing letter. ShipperAG fits tech and product teams shipping their own product who need evidence for a release, and it only tests the web, in Chromium.
Did Octomind shut down?
Yes. In its own archived closing letter, published in spring 2026, Octomind said it was closing after more than three years because it had not found the market validation it needed. The letter said the product would be turned off at the end of May and the company would wind down by the end of June.
The public record matches. Octomind's GitHub repositories, including its CLI, its local debugging tool Debugtopus and its documentation, are archived and read-only, and when we checked on 19 September 2026 the octomind.dev website no longer loaded. The open-source agent runtime at octomind.run is a different product from a different company.
What Octomind did
Octomind was an AI end-to-end testing platform for web apps. Its AI agents explored an app, found user flows and turned them into tests. Each test was stored as a chain of interactions and converted into standard Playwright code just before it ran, and Octomind said it did not use AI at run time.
Tests ran in Octomind's cloud on demand, on a schedule or against pull requests, with results posted back as pull-request comments. Later additions included a recorder, test creation from AI coding tools over MCP, an AI auto-fix that proposed repairs to broken tests for a person to approve, and a command-line tool.
How to migrate from Octomind: what to keep and clean up
The service can no longer export anything, so start with an inventory of what you already hold.
- Keep any Playwright code you saved. Octomind's CLI could print a test's Playwright code, and its debug command could write the Playwright config and test files to disk so they ran with
npx playwright test. - Keep exported YAML test cases. The CLI could pull test cases as YAML, and a December 2025 update added YAML export for a whole test target. Even if you never run them again, they record which flows you covered.
- Rebuild the list if you saved nothing. Octomind's pull-request comments listed each test by name, so old pull requests show what was being tested.
- Remove Octomind from your pipelines. Delete the GitHub Action, Azure DevOps task or scripted trigger, so your builds no longer depend on a service that has gone.
- Revoke access. Delete API keys, remove Octomind's permissions from your repositories, and retire the test accounts and credentials you created for it.
- Close network openings. Remove Octomind's IP addresses from firewall allowlists and shut down any private location worker you ran inside your network.
Octomind alternatives: three routes
Former Octomind users have three realistic routes. The right one depends on whether you need a test suite, a tool that builds one for you, or release evidence that someone else approves.
- Run the Playwright code yourself. Playwright is an open-source testing framework from Microsoft, released under the Apache 2.0 licence, and it runs on any CI provider. You gain full control and take on all the maintenance, including the logins, test data and environments Octomind handled for you.
- Move to another AI testing platform. testRigor turns plain-English test cases into automated tests, and Rainforest QA offers no-code tests with AI self-healing and optional human testers. We compare them in our testRigor and Rainforest QA guides, and cover QA Wolf, Momentic and mabl in all alternatives and comparisons.
- Test for sign-off instead. For many product teams, the job is less "keep a suite green" and more "show the people who approve a release that it does what the requirements say". That is the job ShipperAG is designed for.
Octomind alternatives compared
Here is how the routes compare with what Octomind offered, as of September 2026.
| Aspect | Octomind | Self-run Playwright | ShipperAG |
|---|---|---|---|
| Approach | AI agents found user flows and generated Playwright tests | Test code your team writes and runs | AI QA specialists check each release against your requirements, design system and business rules |
| Who writes and maintains tests | Octomind's AI, a recorder or your prompts; a person approved auto-fixes | Your developers | No suite to write: the coordinator picks the specialists each change needs, within your limits |
| Where it runs | Octomind's cloud | Your machines and CI | Designed for preview and staging builds, and to run inside CI pipelines |
| Evidence and sign-off | Test reports, pull-request comments and Playwright traces | Playwright's HTML report and traces | Verified, Issue found, Needs input and Not covered, in an evidence ledger, plus a self-contained report to sign or share with stakeholders |
| What you keep | Whatever you exported before the shutdown | Your code | The evidence ledger and a self-contained HTML report for each release |
| Pricing (as published) | No longer sold | Free and open source; you pay for CI and your time | From $149 a month, after a free 45-day trial |
| Availability | Closed in 2026 | Available | Private pilot (waitlist) |
Replacing Octomind before your next release? Start with a free 45-day trial, no card.
Work email only. We keep your email, team size and the page you joined from, only to contact you about the ShipperAG pilot. No spam.
What to look for in an Octomind replacement
A shutdown is a good moment to ask harder questions than "does it generate tests?" Ask every vendor, including us:
- Portability. If the vendor changed course tomorrow, what would you keep: code, test cases, reports?
- Who does the work. Who writes the tests, who approves AI changes, and who fixes flaky or broken tests when the product changes?
- What a pass means. Can you see the check that ran? Are failures reproduced before you see them? Does the report say what was not covered, clearly enough for whoever approves the release?
- Access and safety. Can it reach private staging builds, and does it stay inside the hosts you allow?
- Separation. If you test several products or environments, are their context, runs and reports kept apart?
- Commercial fit. Do the pricing model and contract term suit how often your team releases and how many products you test?
Where ShipperAG is different
ShipperAG is not a like-for-like Octomind replacement. Octomind built and ran a test suite for your product; ShipperAG checks each release against what it is supposed to do and hands your team evidence for the release decision. Here is how it works:
- Your requirements are the test plan. Connect a preview or staging URL and add your requirements, design system or brand tokens, business rules and API docs. The coordinator reads each change and picks the specialists it needs.
- Real checks, re-checked. Specialists use a real browser, HTTP and API calls, automated accessibility rules and visual comparison against design tokens. An independent verifier re-runs every finding in a fresh session, and a model's opinion alone never marks anything verified.
- Four plain states. Every item ends as Verified, Issue found, Needs input (the requirements were unclear, so it asks instead of guessing) or Not covered.
- Evidence you keep. Results are sealed in a hash-chained evidence ledger. The report is a self-contained HTML file listing what was verified, issues, open questions and what was not covered, for your team to sign off or share with stakeholders.
- One workspace per team. Projects live inside your team's workspace, each keeping its own context, builds, evidence and reports.
When an Octomind-style tool is the better choice
Be honest about what you valued in Octomind. If it was the suite itself, ShipperAG is probably not your answer.
- You want to own test code. If portable Playwright code was the point, keep running your exported tests or choose a tool that writes code you can take away.
- Your key flows rarely change. A stable suite that runs on every commit pays back over years of releases.
- You need coverage back this week. ShipperAG is a private pilot with a waitlist. If your pipeline has had no end-to-end tests since May, restore them first.
- You need fast, repeatable regression runs. Scripted tests are quicker and more deterministic than agentic checks. We compare the two in agentic QA vs test automation.
Many teams will want both: a small Playwright suite for critical flows, plus release checks against the requirements when a product owner or stakeholder has to sign off.
What ShipperAG does not do
- It is not generally available or self-serve. It is a private pilot with a waitlist: teams start with a free 45-day trial with up to 20 release checks, and paid plans start at $149 a month.
- It is not a like-for-like replacement for a generated regression suite. It is designed to run alongside one.
- It only tests targets you own or have written permission to test, and it stays inside the hosts you allow.
Sources, checked 19 and 20 September 2026: Octomind's own letter, “A letter to our users, customers and readers” (Internet Archive copy, 19 May 2026: “our product will be turned off at the end of May” and “The company will be closed by the end of June”); Octomind's LinkedIn company page (closure post, spring 2026); Octomind's GitHub organisation and profile README; the archived documentation repository, including the FAQ, under the hood, test reports, AI auto-fix, CLI commands, changelog, compliance and security and private location worker pages; the archived Debugtopus and CLI repositories; Octomind's Show HN launch post (September 2023); Playwright's GitHub repository and CI guide; octomind.run. Competitor details change; check their websites for the latest.