A curated record
Two layers: atomic actions — faithful, ungameable ground truth, never rewritten — and a curated narrative the agent scrubs from the raw session. Trustworthy and worth reading.
Keep up with your AI-generated code.
The old code-review model is out. Agents generate too much code, too quickly, for diff-by-diff review to keep pace.
ponens makes each agent session reviewable at a higher level — a curated, verifiable reasoning trace of what was built and why, graded and governed by Computable Governance. So your quality bar and your governance keep pace with the code.
The session — a log
you make capture idempotent
ai · read payment.py
ai · edit payment.py
you yes
ai · pytest → 24 passed
you ok now handle refunds too
ai · edit payment.py
ai · git commit -m "fix"
A reasoning record
Understand the capture path done
Prove capture can't double-charge proved
Implement & test done
assumes a stable idempotency key unverified
Grade B · lineage 100% · policies 4/4 ✓
Make the reasoning rigorous
Separates what was established (proved, tested, verified) from what was merely asserted — and points at where formal methods and other tools belong.
Computable Governance
Policies are machine-checkable rules over the trace — conformance checking (temporal logic over finite traces, à la process mining) brought to AI agents. A real gate, run offline.
From session to review
Emit a trace, grade it, read it in the viewer — zoom from intent-named steps down to atomic actions — and post it on the pull request, where reviewers already are.
Quickstart
ponens agent # the whole workflow, in one command
ponens emit -o trace.json # capture the session as a trace
ponens trace meta set trace.json m3 --title "…" # curate the narrative
ponens trace grade trace.json # the quality rubric
ponens trace check trace.json # Computable Governance — best-practice policies
ponens trace view trace.json # read the reasoning, zoomable