Feature · Traceability
Requirements traceability for AI coding agents links each requirement to the signed test records your agents produce as they work. Microfilm builds that traceability matrix continuously from the evidence your agents write — no spreadsheet to maintain — and exports it as an audit package on demand.
Requirements traceability in Microfilm is a lens over the signed evidence your agents already write — each requirement linked to the signed records that prove it was tested. Nothing to maintain in a spreadsheet: the view is built from the records as they land, and exports as an audit package on demand.
Availability · Team and up — traceability views over your signed evidence are a cloud feature.
What Microfilm is · Microfilm is the evidence engine for teams and agents: a headless MCP server that AI coding agents write signed test evidence to as they work, paired with a web portal where teams search, trace, and share that evidence. Every event is signed with a KMS-backed key and chained, so any record can be verified offline.
The problem
The traditional path is a scramble: chase down what was tested, hunt for proof, and rebuild a traceability matrix by hand in a spreadsheet that’s stale the moment it’s finished. When evidence is already signed and structured, that whole exercise should be a query — not a project.
How it works
Each requirement is linked to the signed records that exercise it, drawn from the evidence in the portal as it lands. The view is a by-product of agents writing evidence — not a document someone maintains.
Because the links are live, gaps are visible: a requirement with no record, a work item never run. You find the hole before the auditor does.
When an auditor or customer asks, produce a structured package — the requirement-to-evidence view plus the signed records behind it — without reconstructing history.
Every record the view points to is signed and chained, so the package you hand over isn’t just a matrix — it’s a matrix backed by evidence the recipient can verify offline.
Where it fits
Traceability is what makes the discipline upstream worth it. It draws directly on the signed records in the evidence portal, which in turn come from the events agents write through the MCP server — the same evidence, seen requirement-first.
FAQ
It maps each requirement to the signed test records that prove it was tested, so you can show every requirement was exercised and see any that weren’t. It’s a view over your signed evidence, built continuously as agents write evidence — not a separate matrix you maintain.
Yes — that’s what it’s built for. As your AI coding agents test, they write signed records to the Microfilm MCP server, and the traceability view links each requirement to the records proving it was tested. The matrix builds itself from what your agents do, so coverage stays current without anyone maintaining a spreadsheet.
The requirement-to-evidence view plus the signed, chained records behind each link — what was tested, the result, and the signature — so the recipient can verify the package offline.
Yes. Because requirements link live to the signed evidence, an untested requirement shows up as a visible gap in the view rather than a surprise at the review.
No. Traceability views and audit-package export are cloud capabilities on the Team plan and up. Writing and signing evidence through the MCP server is free on every plan.
Create an account, connect your agent to the MCP server, and write your first signed record.