Verification contracts
Author verification contracts so agent and CI work can be checked against stated outcomes — structured ACs that humans and automation can both evaluate.
FNTM · Coming soon
FNTM QA
Evidence-backed verification across every surface
Author contracts, walk real UIs mid-chat, and stamp approve or reject on signed evidence — web tabs, Electron apps, and iOS/Android simulators under one FNTM QA Manual MCP plane.
Coming soon — join the waitlist for launch updates
What FNTM QA covers
Evidence review
Review runs and evidence with a blind-first queue so humans can stamp approve or reject on real artifacts. The product dashboard for evidence lives on qa.fntm.ai.
Native surface Manual MCP
Mid-chat attach to Electron desktops (CDP), iOS Simulator, and Android Emulator with the same observe / act / screenshot verbs used for web — so agents dogfood BRWSR, PWR, and mobile the way they already dogfood browser tabs. Web Browser MCP stays for IDE web; Detox and Playwright Electron remain CI ship gates.
Dual-backend browser + BRWSR path
Interactive web walkthroughs run through a Law-3 harness on FNTM BRWSR Sessions or Playwright overflow — agents never own the browser or the evidence signature. BRWSR is the agentic browser path as the platform matures.
Stamped native evidence
Desktop and mobile Manual MCP runs produce stamped evidence under native/electron, native/ios, and native/android — chain-of-custody packets reviewers can audit alongside web traces.
Shadow runs & stamps
Enqueue shadow runs and print verdicts so teams can see posture before stamps gate release workflows. Join the waitlist with productId fntmQa.
Join the waitlist
Marketing primary is fntm.ai/products/qa. Product host qa.fntm.ai is the evidence dashboard — not the marketing home for this line.
Plan-sourced capability inventory
Coming soon — fntm-qa-agent-ci-cli (Agent/CI CLI) Coming soon — fntm-qa-blind-first-human-evidence-review (Blind-first human evidence review) Coming soon — fntm-qa-bounded-fix-loop (Bounded fix loop)