Purpose
Gibs MCP Server is used to provide regulatory compliance support for AI-driven development tools, connecting AI development environments directly to a regulatory compliance knowledge base. It currently covers three major regulations: the AI Act (EU Artificial Intelligence Act), GDPR (General Data Protection Regulation), and DORA (Digital Operational Resilience Act), involving 240+ articles. All responses include article-level citations to binding legal texts, and the corpus is updated in real time as regulations evolve.
Main Capabilities
The tools provide the following features:
classify_ai_system: Classifies an AI system according to AI Act risk levels (unacceptable, high-risk, limited-risk, minimal-risk) and provides complete legal reasoning. The returned content includes: risk classification (for example, high-risk under Annex III, point 1(a)), the specific Article 6 analysis and cross-references to Annex III, applicable obligations based on the classification, and relevant exemptions or exceptions (if any). Example prompt: "Classify my facial recognition attendance system for schools".
check_compliance: Ask any regulatory compliance question and receive a structured answer with article-level citations. Example: asking about GDPR data breach notification requirements returns the obligations under Article 33 (notify the supervisory authority) and Article 34 (inform data subjects), including the 72-hour notification deadline, required content, and exceptions; asking whether DORA applies to a fintech startup handling payments returns an Article 2 scope analysis, including financial entity classification, applicable ICT risk management obligations, and relevant delegated act requirements.
There is also a parameter-free health check tool used to confirm whether the compliance API is available.
How to Use
- Get an API Key: register at gibs.dev and obtain an API key from the dashboard.
- Connect via Claude Desktop: add the configuration to the Claude Desktop configuration file (claude_desktop_config.json).
- Authentication: when calling any tool, pass the API key as the user_api_key parameter, or set it in the environment.
Connection details: any MCP-compatible client can connect to the Gibs server; tested clients include those supporting MCP SSE transport. Python SDK installation command: pip install gibs.
Applicable Scenarios
Suitable for developers and compliance personnel who need to assess EU regulatory compliance for AI systems, especially in scenarios such as AI system risk classification, GDPR data breach obligation inquiries, and DORA scope analysis. It can be used to integrate compliance checks into AI development tools.
Limitations
The repository materials do not provide pricing, performance, or effectiveness data. Specific API invocation methods, detailed configuration, the complete tool list, and other information are subject to the official documentation (docs.gibs.dev) and API reference (docs.gibs.dev/api-reference). The project uses the MIT license and is built by Gibbr AB.
