apronagents
apronagents
Apron Agents splits a task across agents that each work in an isolated, disposable sandbox git — then merges their work one reviewed chunk at a time. Your real remote is never touched.
Apron Agents
merge cockpit
supervised
Plenty of tools run agents. Apron is about what happens between the agents and your codebase.
On launch, Apron creates a bare repo in a temp dir — a fully local "fake GitHub." Every worker codes in its own clone and pushes branches there. One audited module is the only path to git, so nothing half-baked can ever reach your real remote.
A live air-traffic-control view: watch the planner think the moment you dispatch, see every file each agent is touching on a live workspace map, then review chunk-by-chunk. Small, focused diffs; approve or send back with a reason; merges land one at a time with tests on every candidate. Autonomous mode is the same machinery — merges on green.
./run start sets up the environment, boots everything, and opens the dashboard. When the task completes, the merged result is copied into your working directory — the only bridge to reality — and the tool stops.
Workers are pluggable. Use your Claude Code subscription, your ChatGPT plan, a raw API key — or try the whole flow with no account at all.
| runner | powered by | needs |
|---|---|---|
claude-code | the claude CLI, headless | any Claude plan — whatever Claude Code already uses |
codex | the codex CLI, headless | a ChatGPT plan or OpenAI key |
api | the Anthropic API directly | ANTHROPIC_API_KEY |
demo | fake in-process agents | nothing — kick the tires account-free |
Agents themselves are editable markdown files, not code — Apron discovers your existing .claude/agents/ definitions read-only and hot-reloads edits from its own overlay.