Initial public release
First stable release of the Complior open-source compliance platform.Scan Engine- 5-layer static analysis (L1–L4 deterministic + L5 LLM opt-in)
- 57+ AI framework detection (LangChain, CrewAI, OpenAI, Anthropic, Vercel AI, HuggingFace)
- 37 secret patterns (API keys, tokens, passwords)
- 33 code pattern rules across 8 compliance categories
- Cross-layer verification (7 consistency rules)
- SARIF v2.1.0 and JSON output
- 680 dynamic probes across 6 categories (bias, transparency, security, conformity, accuracy, robustness)
- Deterministic + LLM-judged probe modes
- Custom request template support for any API format
- Adversarial red-team testing
- 36-field standardized AI system identity card
- Mode 1 (Auto): 65–70% auto-filled from code analysis
- Ed25519 cryptographic signing
- Export: A2A, AIUC-1, NIST formats
- Evidence chain with SHA-256 hash linking
- 18 remediation strategies (code, document, config)
- Preview with score impact prediction
- AI-assisted document generation (opt-in)
- Full undo capability
- 40+ headless commands
- 9-page interactive TUI dashboard
- Background daemon with file watcher (200ms compliance gate)
- 5 color themes
@complior/sdkruntime compliance middleware- 14 hooks (pre + post)
- 4 provider adapters (OpenAI, Anthropic, Google, Vercel AI)
- PII sanitization (50+ patterns, 6 categories)
- Prohibited practice detection (138 patterns)
- 8 MCP tools for AI agent integration
- Guard tools calling SDK hooks
- 108 EU AI Act obligations mapped
- AIUC-1 (15 requirements)
- Multi-framework scoring (EU AI Act + AIUC-1)
Beta release
- S05 SDK Production complete (30/34 user stories)
- 1,691 tests passing (862 TS + 414 SDK + 415 Rust)
- Compliance diff feature
- Multi-agent awareness
- Cost estimator
- Industry-specific patterns (HR, finance, healthcare, education, legal)