62304 §5.2.2The standard expects a hazard-controlled requirement set; the specification does not expose…
40NEW Deterministic clause mapping for IEC 62304 · 21 CFR 820.30 · ISO 26262 · DO-178C · SOC 2 · ISO 13485
Medical-device SRS audit preparation
Find the traceability gaps before the audit.
TraceForge reads your specification locally, identifies unclear and unverifiable requirements, generates linked verification cases, and exports an evidence report your quality team can review before submission.
● Local parsing · nothing uploaded● Rule-based, reproducible scoring● CSV + Markdown evidence packs
Live engine outputSAMPLE SRS
85readiness
IEC 62304 · grade A
1 major · 0 minor · 1 observationsEvidence is reviewable preparation material. Human review is still required before relying on it for an audit.62304 §9 / ISO 14971Dependencies without a documented degraded path are the most common source of post-release…
608 clauses assessed8 req → 24 tests
5.7hmanual effort to build this one matrix by hand
$827at a typical $145/h compliance consultant rate
8clauses scored per document, per standard
0documents sent to a server during analysis
Pick the standard you have to satisfy
The four moves
From specification to submission in one sitting.
- 01Drop the SRS
PDF, DOCX, TXT or Markdown parsed in your browser. Requirement IDs, source lines, and wording quality are recovered exactly.
- 02Map to clauses
Each requirement is attached to the clause it satisfies, and each generated case becomes the verification method for that clause.
- 03Read the findings
Ambiguous wording, unmeasurable promises, untraced requirements and missing failure paths are written the way an auditor would write them.
- 04Export the pack
Clause assessment, bidirectional matrix, remediation list, statement of coverage, and sign-off block — Markdown plus CSV.
For teams with an audit date
Buy back the week before the finding.
A traceability finding is not a text problem. It can delay a submission, reopen a review, or consume the quality team for weeks. Run the free scan first; request a $490 human-reviewed gap report when the risk is real.
Observed
Source text, extracted requirement IDs, linked verification cases, clause status. TraceForge can prove this now, from your file.
Computed
Readiness score, finding severity, manual-hour estimate. Deterministic rules — inspectable, not a model’s opinion.
Not claimed
Code coverage, CI results, real audit outcomes. TraceForge does not assert what it cannot see.
Does my specification leave this machine?
No. Text extraction and analysis run in the browser. Only a project summary — name, counts, coverage — is registered when you choose to keep dashboard history.
Is this legal or regulatory advice?
It is engineering evidence. Clause mapping is rule-based and every assessment shows the measurement behind it, so your quality and regulatory owners review the same numbers rather than a black-box verdict.
Why not ask a general AI chat to write the matrix?
Because an audit needs reproducibility. Same document in, same findings out, with the rule that produced each line. Chat answers change and cannot be defended in a design review.
What happens between scans?
Nothing silently. You re-run when the specification changes and compare readiness deltas — that is the product. Unversioned evidence is why most matrices rot.
Start where the risk is