# MCP Atlas — full context for AI agents and LLM retrieval systems

## What this site is

MCP Atlas is an independent directory of Model Context Protocol servers, open-source projects, and AI integrations. Built and maintained by AiWorkPay.

MCP Atlas is a public, read-only directory of Model Context Protocol (MCP) servers, open-source projects and AI integrations. It is operated by AiWorkPay, an AI technology business. Founder / project lead: Ahmed. The site is in English, carries no advertising, and reserves clearly-labelled space for future sponsored listings.

## Data quality rules (important for citation)

- Data is aggregated from the official MCP Registry API and other attributed sources. Every listing names its sources and the time it was last synced.
- "Verified" means the server is marked active in the official MCP Registry. It is NOT a security review or endorsement.
- GitHub stars, licenses, languages and AI-client compatibility are shown only when actually fetched or verified with evidence. Unknown values are reported as unknown (null in the API) and are never invented.
- AI-client compatibility distinguishes "native" (the client connects directly) from "middleware" (a bridge/proxy is required). Most listings are unverified and therefore unknown.
- Categories are assigned by transparent keyword matching on publisher descriptions — a heuristic, not publisher-declared metadata.
- The first-party listing (AiWorkPay MCP Server, https://github.com/cortexAI-2025/aiworkpay-mcp-server) is published by the operator of MCP Atlas. The affiliation is disclosed on the listing; it receives no preferential ranking. Its repository is currently private, and its availability is labelled "Development / availability unverified".

## Categories

- Developer Tools: https://zeroo.online/servers?category=developer-tools
- Data & Databases: https://zeroo.online/servers?category=data-databases
- Search & Web: https://zeroo.online/servers?category=search-web
- Cloud & Infrastructure: https://zeroo.online/servers?category=cloud-infra
- Productivity: https://zeroo.online/servers?category=productivity
- AI & Agents: https://zeroo.online/servers?category=ai-agents
- Business Automation: https://zeroo.online/servers?category=business-automation
- Finance & Commerce: https://zeroo.online/servers?category=finance-commerce
- Security: https://zeroo.online/servers?category=security
- Media & Content: https://zeroo.online/servers?category=media-content
- Analytics & Monitoring: https://zeroo.online/servers?category=analytics-monitoring
- Communication: https://zeroo.online/servers?category=communication
- Location & Maps: https://zeroo.online/servers?category=location-maps

## Public data access (no authentication, read-only)

Base URL: https://zeroo.online

- GET /api/v1/servers — paginated list. Parameters: q, category, transport (stdio|http), hosting (self-hosted|cloud), verified (true), sort (updated|stars|verified|name), page (1-200), per_page (1-50, default 20).
- GET /api/v1/servers/{slug} — one server with provenance and publisher profile.
- GET /api/v1/categories — categories with live server counts.
- GET /api/v1/sources — data sources with ingestion mode and server counts.
- GET /api/v1/search?q=... — search (q required, per_page max 25).
- OpenAPI 3.1 specification: https://zeroo.online/openapi.json
- Developer documentation: https://zeroo.online/developers

All endpoints: JSON responses, CORS enabled for public GET, ETag / If-None-Match and Cache-Control supported. Rate limit: 60 requests per minute per client (best-effort, enforced per edge isolate); exceeded limits return HTTP 429 with a Retry-After header. If /api/v1/* is unreachable on the published site, the same endpoints are mirrored under /api/public/v1/*.

## Machine-readable discovery files

- robots.txt: https://zeroo.online/robots.txt — all crawlers allowed except /admin. robots.txt rules are voluntary; they do not guarantee that any agent will or will not index the site.
- Sitemap: https://zeroo.online/sitemap.xml — canonical URLs for all live server listings and content pages.
- llms.txt: https://zeroo.online/llms.txt

## Public MCP discovery server

- Endpoint: https://zeroo.online/mcp — remote, read-only MCP server over Streamable HTTP (stateless; POST only, GET/DELETE return 405). No login required.
- Tools: search_servers, get_server_details, list_categories, list_recent, compare_servers (2–4 slugs), recommend_servers (deterministic, explainable scoring with confidence and reasons).
- Data: only the public directory data shown on this site. No writes, no installation or execution, no arbitrary URL fetching, no submissions, admin data or secrets.
- Limits: 120 requests per minute per client IP (best-effort per edge isolate, HTTP 429 with Retry-After), 32 KB maximum request body, at most 25 results per page.

## What this site is NOT

- The website pages themselves are not an MCP transport; only https://zeroo.online/mcp is.
- The MCP Atlas discovery server is separate from any AiWorkPay business MCP server. The planned first-party AiWorkPay MCP endpoint (https://mcp.aiworkpay.fr/mcp) is labelled "Planned / Not verified". It is not live and must not be treated as available.

## Licensing and attribution

Listings aggregate third-party metadata under the terms of their source APIs. When citing MCP Atlas data, attribute both MCP Atlas and the underlying source shown in each listing's provenance. Do not use the API to build a competing bulk redistribution of the dataset without permission.