Zero AI keys stored

Ship code while you sleep.

Embed a widget, connect a GitHub repo. fdbck triages every signal, opens draft PRs with your own agent, and stays out of the way until a decision is needed.

Open signup · no credit card · runs on your machine

From a user's gripe to a reviewed PR

Three steps. The work happens on your machine, with your own agent and your own credentials — fdbck never sees your code or your keys.

01

Embed the widget

One script tag drops a feedback button into your app. It captures the page, route, and a PII-scrubbed snapshot automatically.

02

fdbck triages it

Every report is classified — bug or feature, severity, confidence — and ranked in your inbox. The vague ones wait; the clear ones move.

03

A draft PR appears

Approve an item and your own agent writes the fix on your machine, runs your tests, and opens a draft PR for you to review.

Built to be trusted unattended

Ranked, not a firehose

An action-first inbox: type, severity, and confidence on every row, so you decide in seconds — no triage tax.

Fixes you can trust

Each fix runs in an isolated worktree and can run your test suite before the PR opens. Optional auto-merge for the confident, small bug fixes.

Multi-repo, multi-agent

Connect as many repos as you like. Run Claude Code, Codex, or Gemini — pick the vendor and model per phase.

A draft PR, never a surprise

Nothing ships without you. fdbck opens drafts against an isolated branch, labelled and linked back to the original report.

Your code, your machine

fdbck stores no model keys, and never sees your source.

A small local worker does all the model and git work on your machine, with your own Claude, Codex, or Gemini and your own GitHub credentials. fdbck holds the feedback and the state of your queue — not your code, your history, or your logins.

Your inbox is waiting to clear itself.

fdbck — feedback that fixes itself