Pinned Tweet
Mo is live. The world's first AI QA engineer: point it at your app, hundreds of agents bug bash it, every bug comes back with repro steps, logs and video. $250 in free credits to start: momentic.ai
Introducing Mo, the world's first AI QA engineer
Point it at your app and 100+ agents bug bash it in parallel. Every bug comes back with repro steps, logs, and a video
In our evals vs Codex with Playwright MCP: → 3.8x more bugs → 9x more bugs per hour → half the cost per bug
50 in free credits. Drop your app URL, Mo will bug bash it and DM you the report today
0
0
6
1,678
Momentic retweeted
That's a wrap!
Thank you to everyone who came out to our first SF Software Factory Meetup last week. At over 150 RSVPs, we were extremely pleased with the turnout!
We loved going deep on AI-powered QA with @wuweiweiwu and getting an inside look at Mastra's home-grown software .
See you in November. ⚙️⚡️
Thank you to everyone who came out to our first SF Software Factory Meetup last week. At over 150 RSVPs, we were extremely pleased with the turnout!
We loved going deep on AI-powered QA with @wuweiweiwu and getting an inside look at Mastra's home-grown software .
See you in November. ⚙️⚡️
0
4
18
1,793
Momentic retweeted
Last week, @sazabi hosted San Francisco's first-ever software factory meetup.
"The SF Software Factory Meetup".
In my opening remarks, I explained why I felt compelled to start "SF2M".
It came down to three emotions: fear, pain, and hope.
1- Fear:
I'm terrified by software factories. We're quickly headed towards a world where humans don't write software at all, and I don't know what that means for our profession.
2- Pain:
We've spent the past 40+ years developing best practices for how to build software. Suddenly, AI has thrown all of them out the window.
3- Hope:
Building software has never been more joyful, creative, and fun. Today, a single person can build and operate apps that used to require entire engineering orgs. What amazing things must be possible in this new era?
This is why I started SF2M.
Not because Sazabi has "figured out" software factories, but precisely because we have not. We want to bring together the people experimenting, questioning old assumptions, and figuring out what comes next.
It's clear we're not alone in thinking about these questions. :)
Thank you to everyone who joined us at SF2M vol. 1.
We'll see you in November. 🏭
"The SF Software Factory Meetup".
In my opening remarks, I explained why I felt compelled to start "SF2M".
It came down to three emotions: fear, pain, and hope.
1- Fear:
I'm terrified by software factories. We're quickly headed towards a world where humans don't write software at all, and I don't know what that means for our profession.
2- Pain:
We've spent the past 40+ years developing best practices for how to build software. Suddenly, AI has thrown all of them out the window.
3- Hope:
Building software has never been more joyful, creative, and fun. Today, a single person can build and operate apps that used to require entire engineering orgs. What amazing things must be possible in this new era?
This is why I started SF2M.
Not because Sazabi has "figured out" software factories, but precisely because we have not. We want to bring together the people experimenting, questioning old assumptions, and figuring out what comes next.
It's clear we're not alone in thinking about these questions. :)
Thank you to everyone who joined us at SF2M vol. 1.
We'll see you in November. 🏭
That's a wrap!
Thank you to everyone who came out to our first SF Software Factory Meetup last week. At over 150 RSVPs, we were extremely pleased with the turnout!
We loved going deep on AI-powered QA with @wuweiweiwu and getting an inside look at Mastra's home-grown software .
See you in November. ⚙️⚡️
3
3
18
1,142
Momentic retweeted
I'm becoming more bullish on specialized agents, e.g.:
- @MomenticAI for QA (s/o @wuweiweiwu)
- @greptile for PR reviews (s/o @dakshgup)
- @sazabi for observability/monitoring
- 2027 for Agent Experience (s/o @heymikasagi )
Claude/Codex can do a lot of this stuff reasonably well out-of-the-box but they certainly aren't optimized for it, which makes them expensive and slow.
More importantly, there seems to be a HUGE amount of value in "getting it just right" for a particular domain/use-case.
For example, Sazabi has spent a ton of time thinking about our core "issue" primitive - what the different statuses and state transitions should be, how the issue should appear when it is rendered in Slack, what the action buttons should be and how they are laid out, etc.
I see Momentic, Greptile, and 2027 doing the same thing for their respective domains/use-cases and it really shows!!!
I struggle to articulate this and assume everyone else does too, which could explain why people seem to underestimate/undervalue specialized agents in general
- @MomenticAI for QA (s/o @wuweiweiwu)
- @greptile for PR reviews (s/o @dakshgup)
- @sazabi for observability/monitoring
- 2027 for Agent Experience (s/o @heymikasagi )
Claude/Codex can do a lot of this stuff reasonably well out-of-the-box but they certainly aren't optimized for it, which makes them expensive and slow.
More importantly, there seems to be a HUGE amount of value in "getting it just right" for a particular domain/use-case.
For example, Sazabi has spent a ton of time thinking about our core "issue" primitive - what the different statuses and state transitions should be, how the issue should appear when it is rendered in Slack, what the action buttons should be and how they are laid out, etc.
I see Momentic, Greptile, and 2027 doing the same thing for their respective domains/use-cases and it really shows!!!
I struggle to articulate this and assume everyone else does too, which could explain why people seem to underestimate/undervalue specialized agents in general
21
2
64
3,338
Momentic retweeted
AI AGENTS CAN BUILD YOUR APP. MO CAN HELP YOU FIND OUT WHERE IT BREAKS.
Mo from @MomenticAI explores any app using a URL, a testing brief, and test accounts if needed. No test scripts required.
It reports test cases, verdicts, coverage gaps, and the bugs it finds.
Every bug work logs, detailed reproduction steps, and a video replay.
That part excites me: you get evidence to investigate, beyond a list of suspected issues.
I built llmexplorer.org in 2025 as a fun project and am planning a refresh, so I put it through Mo.
In one single run:
→ Multiple agents explored pages in parallel, including search filters and model listings
→ They tested external links and UI edge cases, completing 55 test cases
→ The audit surfaced 8 reported bugs, including broken links, clipped organization chips, and filter-trigger issues
After fixing a bug, I can use the reproduction steps to write a regression test in Momentic and run it locally or in CI.
Explore → reproduce → fix → add regression coverage.
Mo doesn’t guarantee a bug-free app, but it gave me a concrete starting point for the refresh.
Watch the demo below. Seeing the agents work through my own site and surface actionable findings is pretty great 👀
Try Mo in the first reply 🧵↓
Mo from @MomenticAI explores any app using a URL, a testing brief, and test accounts if needed. No test scripts required.
It reports test cases, verdicts, coverage gaps, and the bugs it finds.
Every bug work logs, detailed reproduction steps, and a video replay.
That part excites me: you get evidence to investigate, beyond a list of suspected issues.
I built llmexplorer.org in 2025 as a fun project and am planning a refresh, so I put it through Mo.
In one single run:
→ Multiple agents explored pages in parallel, including search filters and model listings
→ They tested external links and UI edge cases, completing 55 test cases
→ The audit surfaced 8 reported bugs, including broken links, clipped organization chips, and filter-trigger issues
After fixing a bug, I can use the reproduction steps to write a regression test in Momentic and run it locally or in CI.
Explore → reproduce → fix → add regression coverage.
Mo doesn’t guarantee a bug-free app, but it gave me a concrete starting point for the refresh.
Watch the demo below. Seeing the agents work through my own site and surface actionable findings is pretty great 👀
Try Mo in the first reply 🧵↓
13
28
62
19,797
Momentic retweeted
:0 its me
we just made bug bashing an iOS or Android app a one-prompt job
@_48nes on our team gave Mo, our AI QA engineer, this prompt:
"bug bash this app thoroughly"
it came back with 8 bug reports, and nobody wrote a single test
every bug came with a recording of what Mo tapped and you can confirm it's real before anyone starts on a fix
before a release, most mobile teams find these bugs one of two ways:
1. someone taps through every flow by hand
2. the team maintains end-to-end tests that fail whenever a screen gets rearranged
Mo does the tapping itself:
→ explores the app on an emulator first
→ splits the app into sections
→ hands each section to a subagent
→ all the subagents run in parallel
every subagent gets its own emulator on purpose. when agents share one, they wait on each other and change each other's app state, so adding more agents stops making the run any faster
one emulator per subagent is only practical because our hosted emulators boot in ~1s
we're still onboarding every mobile team ourselves. if you have an iOS or Android app, DM me and we'll get Mo running on your build
0
1
5
149
Momentic retweeted
It has finally clicked for everyone
true e2e verification is the holy grail
& now people r gonna find out what happens when you have 1k playwright tests to maintain 👀
true e2e verification is the holy grail
& now people r gonna find out what happens when you have 1k playwright tests to maintain 👀
IMO the only tests you should have in the age of coding agents are, in order of priority:
1. Full E2E tests. Nothing mocked out at all. Can even be something that runs in prod on test accounts via Playwright or equivalent.
2. Integration tests. Agents make more mistakes as things bleed between data/API boundaries where schemas can drift.
3. Golden tests. This helps ground the code with real examples of data, and can be used as regressions against edge cases.
Unit tests are bloat 99% of the time if you’re using a frontier model since they’re smart enough now to get the implementation (and subsequent iterations) right on the first shot, so they only add bloat at this point.
36
11
674
73,259
Momentic retweeted
@0xrwu True end user verification is the only thing that matters
Sucks to waste tokens building & maintaining playwright so that’s why notion & whatnot runs on @MomenticAI
Sucks to waste tokens building & maintaining playwright so that’s why notion & whatnot runs on @MomenticAI
0
1
3
259
Momentic retweeted
@JorgeCastilloPr It’s crazy that people still use maestro in 2026
Go agentic w @MomenticAI - notion, whatnot, xero has 👀
Gotta catch up!
Go agentic w @MomenticAI - notion, whatnot, xero has 👀
Gotta catch up!
0
1
2
268
Momentic retweeted
+1 to verification engineering, and cool to see more of it going open source
one thing we learned the hard way:
we spent 2.5 years making tests easier to write at @MomenticAI: plain english instead of playwright, a mix of both agentic and deterministic actions, and self-healing
then in a lot of hard convos, teams told us they didn't want yet another test script to write and maintain, even an easier "agentic" one
we'd made the test easier to write, but the team still owned it. whether it's raw playwright or plain english, someone still has to update it every time the product changes
the ask was simpler: just tell us what broke, why, and how to reproduce it
that's what led us to build Mo, our AI QA engineer. point it at your app and 100+ agents bug bash it in parallel. every bug comes back with repro steps, logs, and a video
no one should have to write or maintain tests again
x.com/rauchg/status/…
one thing we learned the hard way:
we spent 2.5 years making tests easier to write at @MomenticAI: plain english instead of playwright, a mix of both agentic and deterministic actions, and self-healing
then in a lot of hard convos, teams told us they didn't want yet another test script to write and maintain, even an easier "agentic" one
we'd made the test easier to write, but the team still owned it. whether it's raw playwright or plain english, someone still has to update it every time the product changes
the ask was simpler: just tell us what broke, why, and how to reproduce it
that's what led us to build Mo, our AI QA engineer. point it at your app and 100+ agents bug bash it in parallel. every bug comes back with repro steps, logs, and a video
no one should have to write or maintain tests again
x.com/rauchg/status/…
The future is verification-engineering.
Proofs, (e2e) tests, benchmarks, linters…
Some tests will be deterministic, some agentic. This looks great.
9
5
29
5,334
Momentic retweeted
Really love to see all of the experimentation in the "verification of code correctness" space - congrats to @wuweiweiwu and the team for launching Mo.
Got flattened but hope to do an interview within the month to learn more for both the O'Reilly book and the @gatherdev community.
Talking of which, if you're Director+ or Staff+ and building agents for your team, join us in SF or NY this month - luma.com/gather-dev
Got flattened but hope to do an interview within the month to learn more for both the O'Reilly book and the @gatherdev community.
Talking of which, if you're Director+ or Staff+ and building agents for your team, join us in SF or NY this month - luma.com/gather-dev
Introducing Mo, the world's first AI QA engineer
Point it at your app and 100+ agents bug bash it in parallel. Every bug comes back with repro steps, logs, and a video
In our evals vs Codex with Playwright MCP: → 3.8x more bugs → 9x more bugs per hour → half the cost per bug
50 in free credits. Drop your app URL, Mo will bug bash it and DM you the report today
1
2
6
232
7 bugs in GitHub, Cal.com, Tally, Uptime Kuma, SurveyMonkey and Gitea found by Mo
Run it on your app: momentic.ai/?utm_source=x&…
Run it on your app: momentic.ai/?utm_source=x&…
ran Mo's "bug bash" on 6 apps you use every day
7 of the bugs it found, with replays. every one reproduced and accepted by a human 👇
1
0
4
481
Momentic retweeted
@wuweiweiwu MO and I’m impressed. Browser use was strong, it reproduced the steps I gave. There’s also video recording, plus a tab for network events, console errors, etc., similar to Playwright UI mode, which is really helpful.
I’ll dig deeper later for more use cases. For now, I think it could also help with re-running failed tests, just by feeding it the failed report.
I’ll dig deeper later for more use cases. For now, I think it could also help with re-running failed tests, just by feeding it the failed report.
1
1
8
490
Let the bug bash begin.
we spent 2 years making test suites easier to build
now you don't have to create one ever again
give Mo a URL + an objective
its browser agents explore the relevant flows and return a bug report with video + repro steps
no script to write, no suite created just because you momentic.ai/mo
2
1
5
1,045
