Momentic · 2026-11-02 · notable
Momentic Mo — QA agents that bug-bash web and mobile apps on their own
In short
Momentic launched Mo, a QA agent that explores web, iOS and Android apps, reports bugs with repro steps and recordings, and tests pull-request previews. On Momentic's own QA Bench it matched 269 bugs per pass versus 86 for the best Codex setup.
Momentic · 2026-11-02 · 4 sources checked
Point Mo at a URL and it plans, explores and bug-bashes your app, then files bugs with a replay.
Key specs
| Qa bench matched bugs (60 min) | 269.3 |
|---|---|
| Hn points | 161 |
What is it?
Mo is Momentic's autonomous QA agent, launched on Hacker News by the YC W24 company on November 2, 2026. Given a URL, it explores the app in a real browser or emulator, reports each bug it reproduces with steps and a recording, and can turn runs into regression tests for every pull request.
How does it work?
Mo's harness builds a graph of the app as it explores and forks large product areas into parallel sub-agents, with specialised agents for planning, finding elements, and reading images and video. A tunnel lets hosted agents reach private sites and local apps, and new PR testing runs against Vercel and Cloudflare deploy previews. Developers start sessions from the `qa` CLI, the app, a PR comment or a coding agent skill.
Why does it matter?
On QA Bench, Momentic's own benchmark of 1,396 labelled bugs across four open-source apps, Mo matched 269.3 bugs per pass in 60 minutes at 84.2% precision, ahead of Codex (86.3), Claude Code (59.0) and Devin (27.0). Momentic built and graded the test with an LLM judge itself, so the numbers are a first-party claim. The cost per matched bug was $2.67, against $0.80 for Codex with GPT-6.1 Sol.
Who is it for?
QA engineers, product teams shipping web and mobile apps
Try it
npx qa start "test https://your-app.com"