Why compliify
Compliance recorded at the moment code is written
Quarterly questionnaires ask people to remember what their systems do. compliify skips the recall step: while the agent writes code, it records what the code actually does - as structured facts, not prose that goes stale the day it is written.
Facts, not prose
Typed, attributable, diffable - every claim carries file:line provenance, so you always know where an answer came from.
learn moreDeterministic evaluation
Same facts + same rules = same verdict. Every evaluation is replayable for auditors, byte for byte.
learn moreFaster reviews
Your codebase already knows the answers
When a reviewer opens a system, they see evidence with provenance - not a developer's memory of what shipped three sprints ago. What the code can't answer becomes an open question, tracked until someone who knows confirms it.
Read the docs{ "version": 1, "system": "support-triage-bot", "facts": [{ "key": "model.provider", "value": "openai", "provenance": "detected", "source": "src/llm.ts:12", "status": "confirmed" }], "open_questions": ["context.legal_basis"]}
The principle
“The record is a byproduct of development - not an audit after it.”
Built for the rules you answer to
- EU AI Act
- DSGVO / GDPR
- ISO 42001
- ISO 27001
- BetrVG §87
- NIS2
The rundown
What compliify does on every merge.
Observe mode - never blocks a build.
Provenance - every fact carries its source.
Art. 30 - the ROPA renders itself.
Drift - risk changes surface on merge, not at audit.
Interview - the agent asks what code can't know.
CI evidence - signed against the commit.
Pricing
Start small, scale with your portfolio.
Team
Roll facts out across your engineering org.
- One organization, unlimited repos
- Console with review queue
- CI ingest + PR comments
Enterprise
Governance across a portfolio of organizations.
- Multiple organizations
- Custom rule packs
- Priority support