memorable
$ memorable recall "fix the failing auth test"
0.866 procedures/89f11bab-fix-failing-auth-test [lexical,semantic]
$ memorable show procedures/89f11bab-fix-failing-auth-test
THE FIX LANDED IN: tests/auth/session.test.ts
Verified last time by: npm test -- auth
A Coding Agent Solves A Hard Bug Once,
Then Starts From Zero When It Returns.
Recall, Replay, Record.
Memorable keeps the path that worked and hands it back before the next run.
The task comes back. Memorable hands the agent the procedure that fixed it last time.
The agent opens the files the procedure names and runs the command that proved the fix. No repo search.
The finished session is stored as a revision. A different approach is kept beside the old one.
memorable
$ memorable recall "fix the failing auth test"
0.866 procedures/89f11bab-fix-failing-auth-test
How It Works For Coding Agents.
A Claude Code or Codex session ends. Memorable keeps the steps that reached the goal and hands them back when the task returns. 19% fewer turns per task in GBrain on Claude Code.
{ 01 }
Hooks you already have
Claude Code and Codex session hooks send the trace at session end. No change to the agent.
One Run Cold, One Run With Memory.
Cold Run
.. ...Tool calls, QM Codex5
Turns, GBrain Claude Code16
With Memorable
.. ...Tool calls, QM Codex3
Turns, GBrain Claude Code13
Quartermaster on Codex. GBrain on Claude Code.
What Memorable Reads.
4 CONFIGURED
~/.claude/projects
Claude Code transcripts, picked up by memorable backfill
~/.codex/sessions
Codex session rollouts, captured turn by turn
OpenCode hook
A hook records each run when OpenCode finishes
gbrain hook
gbrain records the run through its session end hook
