flatbread.io

mx

mx1.efwd.spaceship.net

mx2.efwd.spaceship.net

ns

launch2.spaceship.net

launch1.spaceship.net

caa

txt

v=spf1 include:spf.efwd.spaceship.net ~all

google-site-verification=AsxbXZARj3XA123NApJC40DGZZ6sv222uyt6wPxF0rk

dmarc

v=DMARC1; p=none;

User-Agent: *
Allow: /
Disallow: /api/
Disallow: /me
Disallow: /login
Disallow: /signup
Disallow: /cli-login

Sitemap: https://flatbread.io/sitemap.xml

# Flatbread

> Flatbread turns an agent session into one link: what was asked, every step, what broke, what fixed it and the diff. People read it. Agents pull it and pick up where it ended.

A run page (https://flatbread.io/r/<id>) shows one agent session: the prompts, each step and tool call, the
failures and how the agent recovered, and the files it changed. Everything on it is read from the
session as recorded, with secrets replaced before upload.

Install with `npm install -g flatbread-cli` (the package is `flatbread-cli`; the npm package `flatbread` is
unrelated). Flatbread reads sessions from Claude Code, Codex, Cursor, OpenCode, Google Antigravity, Gemini CLI, Cline and GitHub Copilot CLI, and `flatbread setup` connects Claude Code, Codex, Cursor, Google Antigravity, OpenCode, Gemini CLI, GitHub Copilot CLI and Cline.

## For agents

- [Install and set up Flatbread](https://flatbread.io/install.md): steps an agent can follow
- [Docs](https://flatbread.io/docs): install, share, pull, what is redacted, and the CLI reference
- [How it works](https://flatbread.io/about): the three steps - share, read, pull - and the questions people ask first
- The execution brief of one run, as markdown: https://flatbread.io/api/runs/<id>/capsule

## Runs

- [All public runs](https://flatbread.io/runs)

## Trust

- [Privacy](https://flatbread.io/privacy): what is read, uploaded, stored, and who hosts it
- [Terms](https://flatbread.io/terms)