Skip to content

Repository files navigation

Agent-Friendly Services

English | 简体中文

Where AI agents plug into 76 popular services: docs, APIs, official MCP servers, llms.txt, CLIs. Every link machine-probed weekly; every capability fact backed by official evidence and a date (methodology) — plus measured agent runs: real agents completing real tasks against the live service, independently verified, transcripts included. 🏆 marks the best measured result in a category — held until someone measures better.

Providers Agent-verified Link health Last update Data: CC BY 4.0

🤖 Agents — add the MCP server, or fetch the whole dataset as one JSON file:

# MCP server — tools: search_providers, get_provider, list_categories, get_stats
claude mcp add agent-friendly-services -- npx -y github:Olorinm/agent-friendly-services

# Or skip the server: the same data as one JSON file
curl -s https://raw.githubusercontent.com/Olorinm/agent-friendly-services/main/generated/providers.json

Other MCP clients: command npx, args ["-y", "github:Olorinm/agent-friendly-services"] (details). Repo map: llms.txt · contribution manual: AGENTS.md.

🧑‍💻 Humans — browse below: each name links to the docs, the trailing links are what officially exists (a missing link means "no known URL", not "confirmed absent"). Full fact sheets with evidence and dates: provider details · measured runs & route comparison: agent runs · spreadsheet: matrix.csv.

🏢 Vendors — fix your own entry in one PR with documentation (not marketing) as evidence. New services enter the candidate pool and are promoted only after a measured agent run — claims are tested, not argued (rules).

Capability matrix — all 76 providers at a glance (✓ supported · ◐ partial · ✗ unsupported · — unknown · Agent: verified routes, h=http c=cli m=mcp)

AI Models

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Anthropic 2026-07-07
Cartesia 2026-07-08
Cerebras Inference 2026-07-08
Cohere 2026-07-07
Deepgram 2026-07-07
DeepSeek 2026-07-07
ElevenLabs 2026-07-07
fal.ai 2026-07-08
Fireworks AI 2026-07-08
Gemini API 2026-07-07
Groq 2026-07-07
Hugging Face 2026-07-07
Luma AI (Dream Machine) 2026-07-08
MiniMax 2026-07-08
Mistral AI 2026-07-07
Moonshot AI (Kimi) 2026-07-08
OpenAI 2026-07-07
OpenRouter 2026-07-08
Alibaba Qwen (Model Studio) 2026-07-08
Replicate 2026-07-07
Together AI 2026-07-07
xAI (Grok API) 2026-07-08
Z.ai (GLM) 2026-07-08

Agent Tooling

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Composio 2026-07-07
Mem0 2026-07-08
n8n 2026-07-08
Vapi 2026-07-08
Zapier 2026-07-07

Code Execution

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Browserbase n/a 2026-07-07
E2B n/a 2026-07-07
Modal 2026-07-07
Steel 2026-07-08

Developer Tools

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Atlassian (Jira & Confluence) 2026-07-08
GitHub ✓ h·c·m 2026-07-07
GitLab 2026-07-07
Postman 2026-07-07

Cloud / Hosting

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Cloudflare 2026-07-08
Fly.io 2026-07-08
Netlify 2026-07-08
Railway 2026-07-08
Render 2026-07-08
Vercel 2026-07-07

Databases

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Chroma 2026-07-08
MongoDB Atlas 2026-07-07
Neon 2026-07-07
Pinecone 2026-07-07
Qdrant 2026-07-07
Redis (Redis Cloud) 2026-07-08
Supabase 2026-07-07
Upstash 2026-07-07
Weaviate 2026-07-08

Web Search / Data

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Apify 2026-07-07
Brave Search API 2026-07-07
Exa 2026-07-07
Firecrawl 2026-07-07
Jina AI 2026-07-08
Perplexity API 2026-07-07
SerpApi 2026-07-08
Tavily 2026-07-07

Payments / Billing

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Lemon Squeezy 2026-07-08
Paddle 2026-07-07
Stripe 2026-07-07

Communication

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Discord 2026-07-07
Lark 2026-07-07
Resend 2026-07-07
Slack 2026-07-07
Telegram Bot API 2026-07-07
Twilio 2026-07-07

Productivity / Storage

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Airtable 2026-07-07
Dropbox 2026-07-07
Linear 2026-07-07
Notion 2026-07-07

Observability / Security

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Datadog 2026-07-07
Grafana (Grafana Cloud) 2026-07-08
Sentry 2026-07-07

Commerce / Marketing

Provider MCP llms.txt OpenAPI CLI Sandbox Self-serve Agent Checked
Shopify 2026-07-07

AI Models · Agent Tooling · Code Execution · Developer Tools · Cloud / Hosting · Databases · Web Search / Data · Payments / Billing · Communication · Productivity / Storage · Observability / Security · Commerce / Marketing

AI Models

Agent Tooling

  • Composio — Tool and integration layer for AI agents (hundreds of app connectors with managed auth), with llms.txt and a hosted MCP directory. MCP · llms.txt · all facts →
  • Mem0 — Memory layer for AI agents (hosted platform + open-source), with REST API, llms.txt, and the official OpenMemory MCP server. API · MCP · llms.txt · all facts →
  • n8n — Workflow automation platform with native AI/agent nodes, a public REST API, official hosted MCP server, CLI, and llms.txt; fair-code and self-hostable. API · MCP · llms.txt · CLI · all facts →
  • Vapi — Voice-agent orchestration API (calls, turn-taking, tool use over phone/web) with an official MCP server and an llms.txt that opens with instructions for AI agents. API · MCP · llms.txt · all facts →
  • Zapier — Automation platform bridging 7000+ apps, with llms.txt and an official MCP endpoint that gives agents access to those integrations. MCP · llms.txt · CLI · all facts →

Code Execution

  • Browserbase — Headless browser infrastructure for AI agents and automation, with session APIs and an official MCP server. API · MCP · llms.txt · all facts →
  • E2B — Isolated cloud sandboxes for running AI-generated code, with llms.txt, an official MCP server, and self-serve keys. MCP · llms.txt · CLI · all facts →
  • Modal — Serverless compute for Python with first-class Sandboxes for agent code execution, llms.txt, and an official CLI. API · llms.txt · CLI · all facts →
  • Steel — Cloud browser API for AI agents (sessions, CDP, anti-bot) — open-source and self-hostable, with llms.txt and a free tier. API · llms.txt · all facts →

Developer Tools

Cloud / Hosting

Databases

Web Search / Data

Payments / Billing

Communication

  • Discord — Chat platform with a versioned bot/OAuth2 API, official OpenAPI spec (preview), webhooks, and documented rate limits. API · OpenAPI · all facts →
  • Lark — Collaboration suite (messaging, docs, calendar) with an open platform, llms.txt, an official CLI with 200+ commands and agent skills, and an official OpenAPI MCP server. API · MCP · llms.txt · CLI · all facts →
  • Resend — Email API for developers with test mode, scoped API keys, idempotency support, and an official MCP server. API · MCP · llms.txt · all facts →
  • Slack — Workspace messaging platform with a mature Web API, granular OAuth scopes, an OpenAPI spec, and llms.txt. API · llms.txt · OpenAPI · CLI · all facts →
  • Telegram Bot API — Free bot platform with instant token issuance via BotFather, webhooks, a documented test environment, and a detailed changelog. API · all facts →
  • Twilio — Programmable messaging and voice APIs with test credentials, an OpenAPI spec, llms.txt, and an official CLI. API · MCP · llms.txt · OpenAPI · CLI · all facts →

Productivity / Storage

Observability / Security

Commerce / Marketing

  • Shopify — Commerce platform with versioned GraphQL APIs, llms.txt, official MCP docs, access-scoped tokens, free development stores, and a CLI. API · MCP · llms.txt · CLI · all facts →

Candidate pool

Submitted, not yet verified — listed for transparency only (how the pool works). An entry is promoted into the index above once an agent passes the M1 first-call run against it and its evidence survives review; until then its claims live only in its YAML file.

Candidate Category Submitted by M1 first-call Claims
paas.build Payments / Billing vendor ✓ pass (2/3) · 2026-07-15 yaml
Xquik Web Search / Data vendor ✓ pass (3/3) · 2026-07-15 yaml

Contributing

One fact = one contribution: report a broken link (2 min, issue form) · resolve one of the 485 open unknowns (15 min) · add a provider (1–2 h, inclusion rules). CI validates everything mechanical; humans only review evidence quality. Full guide: docs/contributing.md.

Have a coding agent? Point it at a checkout and paste:

Read AGENTS.md, then resolve one "unknown" check: find official evidence, update
the provider YAML (or leave it unknown if evidence is genuinely missing), run
npm run validate, and open a small PR.

Related

Fern Agent Score (docs-site readiness scoring — we deliberately don't duplicate it) · Cloudflare Agent Readiness · official MCP Registry · llms.txt hub

License

Code: MIT · Data (data/, generated/): CC BY 4.0

About

A community-maintained, evidence-backed directory of service entry points for AI agents: docs, auth, MCP, OpenAPI, llms.txt — probed weekly.

Topics

Resources

Contributing

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages