Bias Detection MCP
This server provides AI-powered bias detection and fairness auditing tools to help assess, audit, and mitigate bias in AI systems.
quick_scan: Describe an AI system in one sentence and get an instant bias risk assessment.detect_bias: Analyze text for demographic bias patterns, stereotyping, and unfair language, with optional focus on specific protected attributes (e.g., race, gender, age).fairness_metrics: Calculate quantitative fairness metrics from prediction data; supports optional ground truth for advanced metrics like equalized odds and calibration.mitigation_recommendations: Get detailed remediation steps for a specific type of AI bias (e.g., selection, measurement, confirmation, historical, aggregation, or representation bias).regulatory_check: Verify bias compliance requirements against major frameworks including EU AI Act Article 10, NIST AI RMF MAP, and UK standards (eu,us_nist,uk, oralljurisdictions).
Bridges with Cisco's OASF/AGNTCY and ANP protocols to offer bias detection and fairness assessment for AI agent networks.
Integrates with Coinbase's x402 protocol to provide bias detection and fairness assessment for AI agents using HTTP 402 payment flows.
Integrates with Google's A2A and AP2 protocols to offer bias detection and fairness assessment capabilities for AI systems.
Bridges with Linux Foundation's A2A and OASF protocols to provide bias detection and fairness assessment tools for AI governance.
Integrates with Stripe's Agentic Commerce Protocol (ACP) to deliver bias detection and fairness assessment for AI agents involved in commerce.
Bias Detection MCP
AI bias detection + fairness assessment MCP
AI bias detection + fairness assessment MCP. Required by EU AI Act Article 10 + NYC Local Law 144. MIT
๐ Quick Start
# Install via pip
pip install bias_detection_mcp
# Or install via Smithery
npx -y @smithery/cli@latest install bias-detection-mcp --client claudeRelated MCP server: ISO 42001 AI MCP
โจ Features
MCP protocol compliant
Easy installation
Well-documented API
Production-ready
Active maintenance
๐ Documentation
๐ก๏ธ Compliance
This MCP server is built with EU AI Act compliance built-in:
โ Article 9 โ Risk Management System
โ Article 13 โ Transparency & Instructions for Use
โ Article 15 โ Bias Detection & Testing
โ Article 26 โ FRIA Support (where applicable)
โ Article 50 โ AI Content Watermarking (where applicable)
Need help getting compliant? Book a free 15-min diagnostic โ
How MEOK charges
MEOK has no subscriptions and no tiers. You pay per job, and only when the agreed artifact has been delivered to you. There is no public price list. See https://meok.ai/pricing.
This repository's source is MIT-licensed and free to self-host.
๐ค Part of the MEOK Ecosystem
This server is part of the MEOK AI Labs ecosystem โ 26 PyPI packages ยท ~16,300 monthly installs.
Domain | Purpose |
EU AI Act compliance marketplace | |
AI safety & monitoring | |
Sovereign AI platform | |
Legacy modernization |
๐ License
MIT ยฉ CSOAI-ORG
Full catalogue + Anthropic Registry verify links: https://meok.ai
Protocol coverage + Universal PAYG
This MCP is part of MEOK's 47-MCP fleet that bridges every active agent-interop protocol and 30+ regulatory frameworks. See the full coverage matrix at meok.ai/protocols.
Agent interop protocols supported (8 live):
โ MCP (Anthropic) โ native
โ A2A (Google + Linux Foundation, absorbed IBM ACP Sept 2025)
โ IBM ACP โ covered via A2A merge
โ Stripe ACP (Agentic Commerce Protocol) โ Q3 bridge via agent-commerce-protocol-mcp
โ AP2 (Google Agent Payments) โ partial via agent-commerce-payments-mcp
โ x402 (Coinbase HTTP 402) โ partial via api.meok.ai gateway
โ OASF / AGNTCY (Cisco Outshift + Linux Foundation) โ Q3 bridge
๐ ANP (Cisco Agent Network) โ watch-list
Pricing options:
Option | Price | Best for |
Self-host (this MCP) | ยฃ0 โ MIT | Devs |
Each tier above the free self-host adds HMAC-signed attestations verifiable at
verify.meok.ai. Linux Foundation governance on the A2A spine means EU regulated
buyers can deploy without vendor-lock-in objections.
Configuration
Add to your claude_desktop_config.json (Claude Desktop) or your MCP client config:
{
"mcpServers": {
"bias-detection-mcp": {
"command": "uvx",
"args": ["bias-detection-mcp"]
}
}
}Or: pip install bias-detection-mcp then run the bias-detection-mcp command (stdio transport).
Examples
Once configured, ask your assistant, for example:
"Use
quick_scanto โฆ""Use
detect_biasto โฆ""Use
fairness_metricsto โฆ"
Available Tools
5 toolsdetect_biasA
Analyze text for demographic bias patterns, stereotyping, and unfair language.
Args: model_output: The AI-generated text to analyze for bias. protected_attributes: Comma-separated list of attributes to check (e.g. "race,gender,age"). Leave empty for auto-detection. api_key: Optional MEOK API key for pro tier.
Behavior: This tool generates structured output without modifying external systems. Output is deterministic for identical inputs. No side effects. Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice.
Inputs: model_output (required), protected_attributes, api_key.
fairness_metricsA
Calculate fairness metrics from prediction data. Input format: comma-separated values with group labels.
Provide predictions as 'group:prediction' pairs separated by commas. Example: "male:1,female:0,male:1,female:1,male:0,female:0"
If ground_truth is provided, use same format for actual outcomes to compute equalized odds and calibration metrics.
Args: predictions: Comma-separated group:prediction pairs (e.g. "male:1,female:0,male:1"). ground_truth: Optional comma-separated group:actual pairs for outcome-based metrics. api_key: Optional MEOK API key for pro tier.
Behavior: This tool is read-only and stateless โ it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice.
Inputs: predictions (required), ground_truth, api_key.
mitigation_recommendationsA
Get detailed remediation steps for a specific type of AI bias.
Args: bias_type: Type of bias to get recommendations for. Options: selection, measurement, confirmation, automation, aggregation, representation, evaluation, historical. api_key: Optional MEOK API key for pro tier.
Behavior: This tool is read-only and stateless โ it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent โ calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: bias_type (required), api_key.
quick_scanA
Describe an AI system in one sentence -> instant bias risk assessment. No API key required.
Behavior: This tool is read-only and stateless โ it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent โ calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: description (required).
regulatory_checkA
Check bias requirements against EU AI Act Article 10 and NIST AI RMF MAP requirements.
Args: jurisdiction: Jurisdiction to check against. Options: eu, us_nist, uk, all. api_key: Optional MEOK API key for pro tier.
Behavior: This tool is read-only and stateless โ it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent โ calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: jurisdiction, api_key.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelogโฆ
TDQS
Scored across 5 tools
Each tool has a distinct purpose: detect_bias for text analysis, fairness_metrics for prediction data, mitigation_recommendations for remediation steps, quick_scan for instant assessment, and regulatory_check for compliance. No overlap; agents can easily differentiate.
All names use snake_case and follow a descriptive prefix + noun pattern. However, detect_bias uses a verb (detect) while others use adjectives (fairness, mitigation, quick, regulatory), creating a minor inconsistency.
With 5 tools covering analysis, metrics, recommendations, quick assessment, and regulatory checks, the set is well-scoped and manageable, neither too sparse nor too heavy for the domain.
The tools cover core bias detection workflows: text analysis, tabular metrics, remediation, and regulatory checks. Missing are features like reporting or multi-modal support, but these are not critical for the stated purpose.
Maintenance
Related MCP Connectors
AI governance MCP server for EU AI Act compliance and jurisdiction verification
MEOK MCP Hardening MCP โ automated security red-team for any MCP server. Maps OWASP LLM Top 10
Agent Identity Trust MCP Server by MEOK AI Labs
AI Incident Database (AIID) MCP
Related MCP Servers
- AlicenseAqualityCmaintenanceEU AI Act Compliance - MCP server providing AI-powered tools and automation by MEOK AI Labs111MIT
- AlicenseAqualityCmaintenanceISO 42001 AI - MCP server providing AI-powered tools and automation by MEOK AI Labs103 npm37 PyPI1MIT
- AlicenseNot gradedqualityCmaintenanceDeepfake Detector - MCP server providing AI-powered tools and automation by MEOK AI Labs8 npm41 PyPIMIT
- AlicenseNot gradedqualityCmaintenanceRisk Assessment AI - MCP server providing AI-powered tools and automation by MEOK AI Labs4 npm37 PyPIMIT