bidda-mcp
Source-verified regulatory and compliance intelligence: 10,000+ obligations across 39 pillars, each grounded in a primary legal source with a content hash. Covers the EU AI Act, GDPR, DORA, NIS2, HIPAA, Basel III and the MITRE ATT&CK/ATLAS families.
README
Bidda Sovereign Intelligence MCP Server
10,000+ source-verified regulatory compliance nodes across 39 sovereign pillars. Built to reduce hallucination by grounding every node in primary legal sources.
Actively maintained. Last reviewed 2026-07-19 - 10,000+ nodes across 39 pillars, 25 MCP tools (server v1.6.0). Live counts always available via list_pillars() and GET https://bidda.com/mcp.
What is Bidda?
Bidda is a sovereign compliance intelligence registry. Every node traces to a primary legal source (avg 7 citations per node) and contains machine-executable deterministic logic, not summaries, not PDFs, not paraphrased commentary.
Pillars covered: EU AI Act · GDPR · NIST AI 600-1 · Basel III · HIPAA · DORA · FATF · SOC 2 · ISO 27001 · CCPA · PIPEDA · APRA CPS 234 · MiCA · POPIA · plus a MITRE layer across ATT&CK Enterprise/Mobile/ICS, D3FEND, ATLAS and CAPEC, and 150+ others across 39 compliance domains.
MCP Endpoint
POST https://bidda.com/mcp
GET https://bidda.com/mcp (server info, open this in a browser to inspect)
Transport: Streamable HTTP (MCP 2025-03-26). No API key required for the discovery tier.
Tools (25)
Free discovery and intelligence (no key required)
| Tool | Description |
|---|---|
list_pillars |
List all 39 compliance pillars with live node counts |
search_nodes |
Search by keyword across 10,000+ nodes; returns title, ID, pillar, and BLUF (plain-language obligation) |
get_node |
Fetch a specific node by ID; returns summary plus link to machine-executable workflow |
get_dependency_chain |
Walk the prerequisite chain for a node (1 to 4 hops). Plan a full compliance posture from one entry node. |
get_crosswalk |
Cross-framework mapping dimensions for a node (e.g. GDPR Art 17 to CCPA right-to-delete to POPIA Sec 24) |
get_latest_changes |
Regulatory change feed: most recently updated nodes, optional pillar filter |
get_jurisdiction_bundle |
All nodes that apply in a specific jurisdiction (EU, US, UK, AU, SG, IN, CA, CN, ZA, JP, BR and others) |
get_mitre_mapping |
MITRE technique ID to Bidda node plus mapped NIST/ISO/PCI/HIPAA/NIS2/DORA controls. Across 6 frameworks. |
check_action_compliance |
Pre-flight runtime check. Describe an intended action in natural language; get ranked applicable regulations plus LOW/MODERATE/HIGH risk indicator. |
browse_topics |
Browse the registry by cross-cutting compliance topic (e.g. data breach notification, AI transparency) across every pillar and jurisdiction |
Subscriber tools (pass your Bidda key as api_key; a free trial counts)
| Tool | Description |
|---|---|
compare_jurisdictions |
Compare how jurisdictions address a topic side by side, including where their numeric thresholds differ (e.g. a breach-notification deadline of 72 hours versus 30 days). Does not rank which is stricter. |
create_attestation |
Create a signed, time-stamped record of which rules a person or AI agent relied on for a decision, with a public verify URL |
point_in_time |
Signed record of which committed version of a rule was authoritative at a past date, anchored to the public history chain |
watch_changes |
Subscribe to email or webhook alerts when a watched rule or pillar's primary source changes |
open_run |
Open a run ledger for a whole task or conversation (for example a support-bot chat). Returns a run_id. |
record_run_entry |
Append one tamper-evident entry to an open run: the rules consulted, the decision, and the user input as text or a private hash |
seal_run |
Seal a run into one signed run receipt covering every entry, with a public verify URL anyone can check against Bidda's public key |
get_run |
Fetch a run and its entries; a sealed run is publicly readable and reports whether its signature is valid |
consult_node |
Governed runs: fetch a full node and record a verified, hash-pinned entry in an open run in one call, so the sealed receipt proves exactly which version of the rule the agent read |
get_audit_pack |
Export a sealed run as one auditor-ready evidence pack: the signed receipt, the entry chain, and an independent integrity self-check |
drift_check |
Check whether a cached compliance snapshot is still current: pass node ids (optionally with version or content hash) and get fresh, drifted, or withdrawn per node |
create_control_attestation |
Signed record mapping one of your internal controls to the Bidda obligations it addresses, with owner, framework, status, and a public verify URL. An audit trail, not a determination of compliance. |
gap_check |
Walk the public dependency graph from the nodes you cover to surface prerequisite obligations you may be missing |
obligation_deltas |
Obligation-level feed of which primary sources changed or were withdrawn since a date, filterable by pillar or node |
oscal_assessment_results |
Export a sealed governed run as a NIST OSCAL assessment-results document for GRC tooling |
The discovery responses for every free tool are free. Full vault unlock (deterministic_workflow, actionable_schema, full primary_citations) costs $0.01 per node via Skyfire JWT or USDC on Base. The subscriber tools require an active Bidda subscription, which includes a free trial.
Quick Start: Claude.ai / Claude Desktop / Cursor / any MCP HTTP client
Point your MCP client at:
https://bidda.com/mcp
That is it: no install, no API key, no config file for the discovery tier. The full server-info manifest is available at GET https://bidda.com/mcp (open in a browser to inspect available tools).
Run it locally (Node or Docker)
This repo also ships a small local stdio MCP server (server.js) that implements the free discovery and intelligence tools by calling the public Bidda REST API - no API key required. Use it when you want the server running on your own machine or CI.
With Node (18+):
npm install
node server.js
With Docker:
docker build -t bidda-mcp .
docker run -i --rm bidda-mcp
Claude Desktop / Cursor config:
{
"mcpServers": {
"bidda": {
"command": "node",
"args": ["/absolute/path/to/bidda-mcp/server.js"]
}
}
}
The local server exposes seven tools: list_pillars, search_nodes, get_node, get_dependency_chain, get_latest_changes, browse_topics, and check_action_compliance. The subscriber, vault, and attestation tools are available only on the hosted endpoint at https://bidda.com/mcp.
Example Queries
list_pillars()
search_nodes("GDPR data breach notification 72 hours")
search_nodes("Basel III capital requirements", pillar="Banking & Global Finance")
get_node("eu-ai-act-article-13-transparency")
get_dependency_chain("nist-csf-2-0-govern", max_depth=3)
get_crosswalk("gdpr-article-17-right-to-erasure")
get_jurisdiction_bundle("singapore", limit=25)
get_mitre_mapping("T1566") # ATT&CK Enterprise (phishing)
get_mitre_mapping("AML.T0020") # ATLAS (AI-specific)
check_action_compliance("process EU resident biometric data", jurisdiction="eu")
browse_topics("data breach notification")
compare_jurisdictions("data breach notification", api_key="YOUR_BIDDA_KEY")
create_attestation(agent="loan-bot-v2", nodes=["gdpr-article-22-automated-decisions"], api_key="YOUR_BIDDA_KEY")
open_run(agent="acme-support-bot", label="chat 8f21", api_key="YOUR_BIDDA_KEY")
record_run_entry(run_id="run_...", nodes=["gdpr-article-17-right-to-erasure"], note="User: delete my data", api_key="YOUR_BIDDA_KEY")
seal_run(run_id="run_...", api_key="YOUR_BIDDA_KEY")
Coverage (live numbers via list_pillars())
| Pillar | Approx. nodes |
|---|---|
| Cybersecurity | ~1,900 |
| Legal & IP Sovereignty | ~700 |
| Banking & Global Finance | ~580 |
| AI Governance & Law | ~570 |
| Medical & Healthcare | ~325 |
| Sustainability & ESG | ~285 |
| Workplace | ~280 |
| + 32 more pillars | (call list_pillars) |
For exact live counts, call list_pillars(). The manifest at GET https://bidda.com/mcp returns the current totals dynamically.
Full coverage browser: bidda.com/intelligence
How Bidda reduces hallucination
Nodes are produced by Bidda's internal deterministic pipeline: source parsers and verification scripts do the heavy lifting, AI assists only in a small, tightly-gated drafting step, and every node passes multiple independent verification gates before publication. The whole design exists to keep drift to a minimum.
Every node has:
- Primary legal citations: real instruments, real URLs, real section numbers (avg 7 per node)
- Deterministic workflow: machine-executable steps, not prose summaries
- Integrity hash: a cryptographic fingerprint of the node content
- Source integrity watcher: regular TLS and content fingerprint checks against live regulator URLs, results published at
/api/v1/registry-health.json - Verbatim mandate on amendments: every regulatory amendment is quote-justified against the source text before it can change a node
No inference without a regulatory anchor. No blog posts. No secondary commentary. No Wikipedia. Bidda is an information tool, not legal advice; review each rule and its primary source before relying on it.
CISA Secure by Design
Bidda has publicly attested to the CISA Secure by Design Pledge, the seven public goals the U.S. Cybersecurity and Infrastructure Security Agency asks software manufacturers to commit to. Additional Bidda and CISA mappings:
- CISA CPG crosswalk: bidirectional mapping between Bidda compliance nodes and CISA's Cybersecurity Performance Goals: bidda.com/cisa/cpg-crosswalk
- CISA Free Tools catalogue: no-cost capabilities Bidda offers federal, SLTT and critical-infrastructure defenders: bidda.com/cisa/free
Registry Health
GET https://bidda.com/api/v1/registry-health.json
Live integrity-check results: source URL liveness, verification coverage percentage, regulatory change detection categories, with a public timestamp on the last sweep.
Links
- Website: bidda.com
- Developer docs: bidda.com/developers
- Full node registry: bidda.com/intelligence
- CISA programs: bidda.com/cisa
- Verify a node: bidda.com/verify
- Methodology: bidda.com/methodology
- Contact: info@bidda.com
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。