Turn every agent session
into a better next run.
Built because we got tired of guessing what actually works.
ax watches every session your coding harness runs, spots the mistakes it repeats, and turns them into small, repo-specific fixes you review and apply — one at a time.
curl -fsSL ax.necmttn.com/install | shruns locally · you review every change · works with the agents you already use
Your sessions, working for the next one.
Every session lands in one local graph your agent can reach back into — to catch a repeat mistake, replay a clean run, recover the moment a test fails, or grade the harness itself. Same loop, whatever you run it in.
Every applied fix moves the score.
Harness Doctor grades how well your setup is working. ax keeps finding small fixes — you apply the ones you like, and the number climbs.
Harness Doctor · this week
ax applied, one review at a time
- axadded
pnpm test to AGENTS.md - axignored
dist/ + .generated/ - axpruned
3 stale CLAUDE.md rules - axadded
recovery_path hook
same harness · four fixes later
HOOK AUTO HARNESS
smoke_test
disable_command
Run ax serve to see yours.
the pipeline · what feeds the graph
auto · pausedFTS median 5.9ms
every event → typed graph → ranked interventions. local SurrealDB at 127.0.0.1. see /features for the schema →six forms: skill · guidance · subagent · hook · automation · harness_check
If it shapes your agent, you should be able to fork it.
ax is AGPL-3.0, local-first, and end-to-end typed. The whole feedback loop lives in files and a database on your machine — inspect it, bend it, fork it. Commercial license available if you need it.
gh repo clone Necmttn/axCloned ax into ./axcd ax && bun install1 284 packages installed in 4.2sax daemon startax dashboard → http://127.0.0.1:8520AGPL-3.0
Free & open · commercial license available
TypeScript
End-to-end, strictly typed
Local-first
SurrealDB on 127.0.0.1
Forkable
Hack the loop, ship a PR