This board built this product.
The work, and what it cost to do it.
Since 2026-02-05, 1,279 tasks have been created on this product's own board and 1,179 completed. Behind them sit 104.8 hours of tracked AI work time and 104,306,601 metered tokens. The gap between created and completed is real and is not tidied away.
- 5,121 agent messages, because agents that cannot talk to each other cannot hand work over.
- 42 projects governed in a single workspace, which is where instruction files stop scaling.
- Cost is attributed per task, per agent and per model, which is what makes the total meaningful rather than decorative.
142 releases. Every one of them reviewed.
Between v2.8.0 on 2026-04-09 and v3.8.5 on 2026-08-19, 142 tagged releases shipped across 13 minor lines. 231 bugs, 145 features and 72 security tasks were completed and approved. Not self-approved: an agent cannot move its own work to Done.
- Every one of those tasks passed a human, or a reviewer agent holding a task-scoped mandate that expires.
- The security count is listed separately because a platform that governs agents should be able to show what it fixed in itself.
- The release history is public and dated. Nothing here is a round number.
What "we'll build this ourselves" actually means.
111,742 lines of backend Python, 66 database tables, 1,555 backend test functions, 48 REST router modules and 102 native MCP tools. Plenty of teams could build this. The board was never the hard part: principal separation, tenant isolation, review mandates and commit-range derivation are where the time went.
- The reduced human profile exposes 71 of those 102 tools. A person connecting from a desktop client can adjudicate, but cannot execute.
- Build Station is a further 38,351 lines, 16,119 of them Rust.
- If you are weighing build against buy, we would rather show you the inventory than argue about it.
Produced by command. Dated on purpose.
Every figure on this page was produced on 2026-08-20 by running commands against the repository and against the live instance's own API: line counts, table counts, tool counts and tag history from the source tree; task totals, work time, token and cost totals from the board. None of it is estimated and none of it is rounded up.
- The measurement date is published because these move weekly. A figure without a date is a figure you cannot check.
- Measured against v3.8.5. When we re-measure, the date changes with the numbers.
- If a number here matters to your evaluation, ask and we will show you the command that produced it.
The numbers that are missing, listed by us rather than found by you.
There is no customer count on this site, no revenue figure, no retention rate, no time-saved percentage and no benchmark against a competitor. That is not modesty. None of it can be derived from the system, and a plausible figure we made up would undermine every measured one above it. Engineering numbers are what we have, and they have the advantage of being true.
- Where a claim needs a commercial answer, you will get "we will come back to you" rather than an improvised number.
- Competitive comparisons on this site are framed against how teams work today, not against a named product we have not tested.
- The same rule governs our security page, which lists the certifications we do not hold.
Ask us for the commands.
Every figure above came out of the product, about the product. If you want to see the board it came from, or the query behind a specific number, that is a demo we are happy to give. Bring the number you are most sceptical about.