FZ-024 · DEVTOOLS
Humantesting
Hail a human tester: a real person on a real device runs your test task.
- Status
- In development
- Pipeline stage
- VALIDATION
- Sector
- DEVTOOLS
- Autonomy target
- NOT SET
- Built with
- 1 live · 2 planned
The problem
An app team has no quick way to get a person to try a build on a real phone and report back what they found.
What it does
Humantesting lets a Factory Zero venture post a test task. A vetted tester in Indonesia runs it on a real device, and the result comes back recorded and checked.
How it works
- A venture posts a test task for its app
- A vetted tester runs it on a real device while it is recorded; an AI reviewer checks the result, and a senior human decides when it is unclear
- The venture gets a verified result, the bug reports filed, and a replayable script
Paid from prepaid credits in USD.
Stop guessing what broke on a phone you do not own.
TodayThe backend is being built, and a staging API is up. The service is not live yet.
Aims
- Autonomous operation
- The AI review and the weekly payout run are meant to run on schedule, and dispatch offers each task to testers as it is posted. Senior reviewers handle the unclear cases and the appeals.
- Intelligence
- An AI reviewer checks each recording against its task, and an AI rewrite turns a tester’s notes into a bug report. A senior human decides when the AI is unsure.
- Growth
- In v1 the clients are Factory Zero ventures only, and the testers are in Indonesia only, with more countries later.
The technical detail
Lets Factory Zero ventures hail a human tester. A test task is posted, dispatched to a vetted tester in Indonesia on a real device, recorded, checked by an AI reviewer (and by a senior human when it is unclear), then returned as a verified result, filed bug reports and a replayable script. Clients pay from prepaid credits in USD, and testers are paid weekly in rupiah. Nothing is built or live yet.
Related
From the same foundation.
- Cratefield. A backend for builders who would rather compile than configure.
- Owlpost. Email API for apps and AI agents, with its own inboxes.
- SupportGenius. Support answers from your docs; anything unanswered goes to your team.
- VibeCaddie. A code review agent for work you shipped without reading every line.