Vaara
Runtime evidence layer for EU AI Act compliance — intercepts agent tool calls, scores risks, and writes audit trails.
Vaara is an open-source runtime evidence layer for AI Act compliance. It intercepts agent tool calls, scores each with a conformal risk interval, and writes hash-chained audit records. The system supports LangChain, CrewAI, and OpenAI Agents SDK through native adapters and generates reports that external auditors can verify without trusting the deployment stack.
Source
Repository: https://github.com/vaaraio/vaara
Maintain Vaara?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[](https://getagentictools.com/mcp/vaaraio-vaara?ref=badge) Related mcp
ScopeBlind MCP security gateway enforcing per-tool policies, rate limits, and approval gates with signed audit receipts.
Actual Budget Integrates with Actual Budget to provide transaction filtering, spending breakdowns by category, monthly summaries with savings ra
Vivado EDA Controls Xilinx Vivado for FPGA development with synthesis, implementation, bitstream generation, and automated diagnostic tools.
Paddle Billing Enables developers to manage billing operations using natural language commands, streamlining financial API interactions for SaaS
PaddleOCR Integrates with PaddleOCR and PP-StructureV3 to extract text from images and PDFs with confidence scores and bounding boxes, plus
Netdata Real-time infrastructure monitoring with metrics, logs, alerts, and ML anomaly detection