Know what your AI coding agents really cost.

Is your AI agent spend worth it? Measure it.

A self-hosted report that puts each agent workflow's cost next to its outcomes, its evidence and the gaps in the data — not just a token count.

Sample report · illustrative dataSeptember · 3 workflows

Cost per merged changetotal spend £10,550

Bug fixes · agent solo£22.55

212 sessions · 141 merged · evidence 88%9 sessions missing CI data

Features · agent + review£76.03

96 sessions · 58 merged · evidence 71%21 sessions missing review data

Refactors · long-running£246.67

41 sessions · 12 merged · evidence 49%wide range: low evidence

Reads Codex and nodewell session records today. Claude Code, Cursor and Copilot: planned.

Outcomes

See what the spend bought.

Cost sits next to outcome and evidence for every session — not just the tokens a vendor dashboard shows.

Comparisons

Test one workflow against another.

Controlled comparisons on the same work, before you switch tools or roll a workflow out.

Honesty

It shows the gaps.

Missing data and uncertainty are on the page, so the report never over-claims a saving.

Send one month of sessions. Get the report free.

After the pilot, planned from £99/month per team. Self-hosted on your own PostgreSQL.

Questions

acoolit is not generally available. Report data shown is illustrative.