Validance¶
Validance is a workflow execution engine for regulated computational environments. It runs containerised tasks defined in JSON or built with the Python SDK, and produces a per-run forensic record suitable as objective evidence under FDA Computer Software Assurance, EudraLex Annex 11, and 21 CFR Part 11.
What Validance produces¶
Every workflow run produces, automatically:
- A run record: workflow definition hash, engine version, triggering user, timestamp, parameter values
- For every task: container image and image hash, command, input file content hashes, output file content hashes, exit code, timing, host
- A complete input-step-output lineage retrievable by API
- A tamper-evident audit trail with built-in verification
Records are produced as a side effect of execution. They do not require additional configuration, instrumentation, or manual reconstruction.
Where to start¶
If you are evaluating Validance against a regulatory checklist, start with Regulatory mapping — what each Validance deliverable corresponds to in Part 11, Annex 11, and FDA CSA terms.
If you are integrating Validance into an existing system, the Concepts section covers workflows, tasks, runs, and audit at the level needed to use the API.
The wire format is documented at api.validance.io/docs. The Python SDK is at github.com/validance-io/sdk-python (Apache 2.0).
Status¶
Validance is closed-source and pre-GA. The Python SDK is open source under Apache 2.0. The engine is licensed and distributed as a binary; self-hosted and air-gapped deployments are supported.
For pilot enquiries, technical briefings, or evaluation access, contact wiktor.lisowski@validance.io.