Waqi - AI Privacy Layer

This is ahosted, commercialMCP server. This repository contains its documentation and MCP Registry manifest — the service itself runs atwaqi.bilazann.com.

AI assistant ──1· tool call──▶ Waqi ──2· read-only fetch──▶ Your tools (Stripe, Xero, CRM…) ▲ │ └──3· redacted data only───┘ (audit log written AFTER redaction — the sensitive values are never stored)

- Redaction first.Names, emails, phone numbers, card numbers, addresses and other personal identifiers are detected and replaced before the response leaves Waqi.
- Provable audit log.Every call is logged — who asked, which tool, how many sensitive items were caught — so you can show clients or regulators exactly what the AI never saw.
- Per-person tokens.Each team member gets their own MCP URL; every call is individually attributed, and access is revoked per person, not per company.
- Read-only by design.Waqi fetches; it never writes to your tools.
- Credentials stay encrypted.Connector keys are stored encrypted and are never visible to anyone — including us.

https://waqi.bilazann.com/api/mcp/YOUR_TOKEN

Claude (web/desktop):Settings → Connectors →Add custom connector→ paste your MCP URL.

claude mcp add --transport http waqi https://waqi.bilazann.com/api/mcp/YOUR_TOKEN

ChatGPT:Settings → Connectors →Advanced→Developer mode→ add the same URL.

Standard MCP configuration(for clients that use a JSON config file):

{ "mcpServers": { "waqi": { "type": "http", "url": "https://waqi.bilazann.com/api/mcp/YOUR_TOKEN" } } }

Tip: in Claude, an attached connector still has to beenabled in the chatyou're using — if tools list but never run, that's the switch to check.

Security and compliance layer for MCP agents. The analyze_prompt tool checks any input — user messages, RAG retrievals, tool outputs — for prompt injection (22 deterministic signatures, 7 languages) and PII before your model sees it. Every verdict returns a signed audit record (SHA-256 + UUID + UTC) you can retain as GDPR Art. 30 evidence. Free tier: 10,000 requests/month.

Cited compliance, legal & security intelligence for AI agents — 300+ law, regulation and standards corpora (GDPR, NIS2, AI Act, DORA) with article-level citations, plus live CVE/KEV/EPSS vulnerability context. OAuth remote server, free tier

Require a named human's offline-verifiable approval before an AI agent takes an irreversible action — payment release, record change, deploy. Two-person rule, Ed25519 Trust Receipts, IETF-drafted, Apache-2.0.

KHEPRA MCP Server smithery badge MCP Registry License Container PQC Sovereign compliance engine with 36,195 STIG/CCI/NIST/CMMC mappings. Air-gappable. Zero token costs. Run ert_scan → get a Godfather Report with dollar-denominated business impact. The only MCP compliance server that runs on your metal — with the World's First DoD PQC STIG built in. PQC-01-STIG-V1R1 — Full Whitepaper → 17 controls covering CNSA 2.0, FIPS 203/204/205, and the NSA's May 2026 MCP security advisory. The world's first DoD-style Post-Quantum Cryptography STIG, including the first PQC controls for agentic AI and MCP deployments.

Six-gate governance for AI agents: PROCEED/PAUSE/HALT decisions with hash-chained audit trails.

Cryptographic runtime governance for AI agents. 20 tools. Sealed policy artifacts, continuous measurement, tamper-evident proof. Ed25519 + SHA-256.

AI Agent Supply Chain Security - Intercepts and validates every package installation, git clone, and script download triggered by AI coding agents before it executes.

Structured AI incident reporting for EU AI Act Article 62 — generates mandatory incident reports, severity classification, root cause analysis, and regulator-ready submissions for serious AI incidents.

Enforces organisational AI usage policies at the agent layer — blocks prohibited model calls, enforces data residency rules, logs policy violations, and ensures AI governance policies are machine-executable.