# AI Agents Accessible APIs

> Catalog of 330 third-party APIs and MCP servers that AI agents can call on a user's behalf after a one-time API key / OAuth / MCP setup. 29 categories. Last checked 2026-09-24; each entry has its own `verified` (last-checked) date and a `link_check`; confirm pricing and task fit in `docs`. Each entry: docs URL, auth method + header hint, MCP server (official/community), free tier, SDKs, OpenAPI and llms.txt links.

Use the JSON files; fields: id, name, category, homepage, docs, auth, auth_hint, mcp{type,url}, free_tier, sdk, openapi, llms_txt, desc_en, notes, verified, plus derived booleans has_free_tier, has_trial, no_auth, stale, link_check{checked_at, docs, mcp} and ratings[] (proof of use: SourceForge rating and review count, GitHub stars, Smithery uses; each with url and fetched_at) and reviews_url (up to 100 user reviews with pros/cons, reviewer names removed; also via the MCP tool get_reviews). Researched on vendor pages, each with source_url and checked_at: free_plan{kind, requires_card, quota, period, watermark}, auth_scheme (OpenAPI securityScheme: type, in, name, scheme, format), base_url, rate_limits, oauth_scopes, async_jobs, data_policy, mcp{maintainer, remote_url, repo, tools, covers_full_api} plus derived mcp.kind (vendor-hosted/official/community/none) and mcp.domain_verified, mcp.config (connection snippet: transport, url or command/args, headers, env, oauth, docs_only; derived: true when built from the endpoint URL alone), mcp.issue. For popular services also call{operation, method, url, example (curl), response_fields, async} and unit_price{summary}, each with source_url and checked_at.

## Data

- [Full catalog (JSON)](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/all.json): all entries in one array
- [Full catalog (text)](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/llms-full.txt): one line per service
- [JSON Schema](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/data/schema.json): entry schema
- Uptime (`uptime` in JSON): our own daily probe of each `base_url` and hosted MCP endpoint from GitHub Actions (US), 30-day window; up = any HTTP status below 500 (MCP: POST initialize); `probes`, `up`, `median_ms` and the last result
- [Operations](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/operations.json): vocabulary of the `operations` field (what an agent can do with a service, e.g. text-to-video, sms, street-geocoding); match your task to these ids
- [SHA256SUMS](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/SHA256SUMS): checksums of llms.txt, llms-full.txt, schema and catalog/*.json (`sha256sum -c SHA256SUMS`)
- [MCP server](https://github.com/eater2/ai_agents_api_library#mcp-server): `npx -y github:eater2/ai_agents_api_library` or remote https://ai-agents-api-library.vercel.app/mcp (Streamable HTTP) — tools search_apis, get_api, get_reviews, list_categories

## Start free

225 of 330 services have a lasting free tier or free usage (JSON field `has_free_tier`; exact limits in `free_tier`), and 58 more give one-off free trial credits (`has_trial`). 21 work with no key at all (`no_auth`), so you can call them immediately without asking the user for credentials:

- SearXNG (self-hosted JSON API) (web-search): Query a self-hosted SearXNG metasearch instance's JSON API to aggregate results from many search engines at once. Docs: https://docs.searxng.org/dev/search_api.html
- Playwright MCP (Microsoft) (web-scraping-browser): Official Microsoft MCP server giving agents structured browser control (click, type, navigate) via Playwright's accessibility tree. Docs: https://github.com/microsoft/playwright-mcp#readme
- arXiv API (knowledge-research): Search arXiv preprints by keyword, author, category or ID and retrieve metadata, abstracts and PDF links. Docs: https://info.arxiv.org/help/api/user-manual.html
- Crossref REST API (knowledge-research): Look up and search scholarly metadata such as DOIs, references, authors, funders and licenses across Crossref-registered works. Docs: https://www.crossref.org/documentation/retrieve-metadata/rest-api/
- GDELT API (knowledge-research): Full-text search global news in about 65 languages and analyze coverage volume, tone and geography in near real time. Docs: https://blog.gdeltproject.org/gdelt-doc-2-0-api-debuts/
- Gitingest API (knowledge-research): Lets an agent turn any public Git repository into a single text digest (summary, directory tree and file contents) sized for an LLM prompt. Docs: https://gitingest.com/llms.txt
- Hacker News API (knowledge-research): Read Hacker News stories, comments, jobs, polls, user profiles and front-page rankings as real-time JSON. Docs: https://github.com/HackerNews/API
- Internet Archive Open Library API (knowledge-research): Lets an agent search books, authors and subjects, fetch work/edition data by ISBN or OLID, search inside book texts and get cover images. Docs: https://openlibrary.org/developers/api
- PubMed E-utilities (NCBI) (knowledge-research): Search, fetch, summarize and cross-link biomedical literature records from PubMed, PMC and other Entrez databases. Docs: https://www.ncbi.nlm.nih.gov/books/NBK25501/
- Wikidata (knowledge-research): Query the Wikidata knowledge graph of entities, facts and relations via SPARQL, the REST API or MCP. Docs: https://www.wikidata.org/wiki/Wikidata:Data_access
- Wikipedia / Wikimedia API (knowledge-research): Search, read and summarize Wikipedia and other Wikimedia content and metadata via REST and Action APIs. Docs: https://www.mediawiki.org/wiki/Wikimedia_APIs
- Kroki (diagrams-software-architecture): Render 25+ diagram-as-code formats (PlantUML, Mermaid, C4, GraphViz, BPMN, etc.) from text via a free HTTP GET/POST API. Docs: https://docs.kroki.io/kroki/
- PlantUML Server API (diagrams-software-architecture): Render UML and architecture-as-code diagrams to PNG/SVG/ASCII via a free public HTTP GET API using encoded diagram text. Docs: https://plantuml.com/server
- Globalping Network Measurement API (developer-devops): Lets an agent run ping, traceroute, MTR, DNS and HTTP tests from thousands of probes worldwide to check reachability, latency and routing of a host. Docs: https://globalping.io/docs/api.globalping.io
- Maven Central Search API (developer-devops): Lets an agent search Maven Central for Java/JVM artifacts by groupId, artifactId, class name or SHA-1 and look up the latest versions before adding a dependency. Docs: https://central.sonatype.org/search/rest-api-guide/
- ntfy.sh Push Notifications API (communication-email-chat): Lets an agent send push notifications to a user's phone or desktop by publishing a plain HTTP POST to an ntfy topic, and read messages back from it. Docs: https://docs.ntfy.sh/publish/
- National Weather Service API (NWS) (maps-geo-weather): Fetch official US government weather forecasts, alerts, and observations via a free public REST API. Docs: https://www.weather.gov/documentation/services-web-api
- Open-Meteo (maps-geo-weather): Fetch weather forecasts, historical weather, and climate data worldwide via a free open REST API, no key required. Docs: https://open-meteo.com/en/docs
- OpenStreetMap Nominatim (maps-geo-weather): Geocode and reverse-geocode addresses from OpenStreetMap data via a free public REST API. Docs: https://nominatim.org/release-docs/latest/api/Overview/
- Overpass API (OSM) (maps-geo-weather): Run custom queries (Overpass QL) over OpenStreetMap data - POIs, roads, boundaries - via a public read-only API. Docs: https://wiki.openstreetmap.org/wiki/Overpass_API
- DexPaprika DEX API (finance-payments): Lets an agent query on-chain DEX data across 36 blockchains: liquidity pools, swaps, token prices, OHLCV history and live price streams. Docs: https://docs.dexpaprika.com/api-reference/introduction

## Categories

Counts per category: services / with free tier / no key needed / official MCP server.

- [Web search](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/web-search.json): 14 services / 9 free tier / 1 no key / 11 official MCP. Search engines and answer APIs that return fresh web results, snippets or full page content in agent-friendly formats.
- [Web scraping and browser automation](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/web-scraping-browser.json): 18 services / 12 free tier / 1 no key / 15 official MCP. Crawl, scrape and convert pages to Markdown/JSON, or drive a remote headless browser (clicks, forms, logins).
- [Knowledge and research data](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/knowledge-research.json): 16 services / 16 free tier / 9 no key / 3 official MCP. Encyclopedias, academic papers, news, computational knowledge and up-to-date library documentation.
- [Image generation and editing](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/image-generation.json): 10 services / 2 free tier / 0 no key / 5 official MCP. Text-to-image, image editing, inpainting, upscaling and background removal.
- [Video generation and editing](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/video-generation.json): 14 services / 2 free tier / 0 no key / 11 official MCP. Text/image-to-video models, talking avatars, programmatic video rendering and video hosting.
- [Speech and audio](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/audio-speech.json): 13 services / 6 free tier / 0 no key / 8 official MCP. Text-to-speech, speech-to-text, voice cloning, dubbing and audio intelligence.
- [Music generation](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/music-generation.json): 6 services / 1 free tier / 0 no key / 1 official MCP. Generate songs, instrumentals, loops and sound effects from prompts.
- [3D generation and assets](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/3d-generation.json): 7 services / 1 free tier / 0 no key / 2 official MCP. Text/image-to-3D mesh generation, texturing and 3D asset libraries.
- [Architecture, CAD and BIM](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/architecture-cad-bim.json): 8 services / 5 free tier / 0 no key / 4 official MCP. Building information models, CAD geometry, construction management and GIS platforms with programmatic access.
- [Diagrams and software architecture](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/diagrams-software-architecture.json): 9 services / 7 free tier / 2 no key / 6 official MCP. Render or edit diagrams (Mermaid, PlantUML, C4), whiteboards and architecture models.
- [Design and UI](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/design-ui.json): 6 services / 4 free tier / 0 no key / 5 official MCP. Read and write design files, generate branded graphics and publish websites.
- [Documents, OCR and presentations](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/presentations-documents.json): 14 services / 8 free tier / 0 no key / 10 official MCP. Parse PDFs and scans, convert file formats, generate slides and collect e-signatures.
- [Translation and language](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/translation-language.json): 6 services / 5 free tier / 0 no key / 3 official MCP. Machine translation, grammar and style checking.
- [Code execution sandboxes](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/code-execution-sandbox.json): 10 services / 5 free tier / 0 no key / 6 official MCP. Isolated cloud environments where an agent can run code, shells and whole dev environments safely.
- [Developer platforms and DevOps](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/developer-devops.json): 12 services / 12 free tier / 2 no key / 11 official MCP. Source control, CI/CD, monitoring, error tracking, incident management and package registries.
- [Cloud and hosting](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/cloud-hosting.json): 10 services / 9 free tier / 0 no key / 10 official MCP. Cloud providers, serverless and edge platforms, and site hosting an agent can deploy to and manage.
- [Databases](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/databases.json): 12 services / 9 free tier / 0 no key / 12 official MCP. Managed SQL, NoSQL, key-value and analytics databases an agent can create, query and manage.
- [Vector search, memory and RAG](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/databases-vector-memory.json): 9 services / 8 free tier / 0 no key / 9 official MCP. Vector databases, retrieval (RAG) services and long-term memory layers for agents.
- [Email, team chat and notifications](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/communication-email-chat.json): 10 services / 9 free tier / 1 no key / 6 official MCP. Send and receive email, post to team chat (Slack, Discord), send push notifications; manage meetings.
- [SMS and messaging apps](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/sms-messaging.json): 10 services / 3 free tier / 0 no key / 8 official MCP. Send and receive SMS and messages on WhatsApp, LINE or Telegram through a messaging provider's API.
- [Voice agents and telephony](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/voice-agents-telephony.json): 8 services / 3 free tier / 0 no key / 7 official MCP. Place and receive phone calls handled by AI voice agents, real-time audio/video rooms.
- [Productivity and workspace](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/productivity-workspace.json): 17 services / 17 free tier / 0 no key / 17 official MCP. Email, calendars, files, notes, tasks and project management.
- [CRM, support and marketing](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/crm-support-marketing.json): 8 services / 6 free tier / 0 no key / 6 official MCP. Contacts, deals, support tickets and email campaigns.
- [Social media](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/social-media.json): 10 services / 7 free tier / 0 no key / 3 official MCP. Read and publish posts, comments and videos on social platforms.
- [Maps, geolocation and weather](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/maps-geo-weather.json): 16 services / 16 free tier / 4 no key / 9 official MCP. Geocoding, routing, places, map tiles and weather forecasts.
- [Finance, payments and market data](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/finance-payments.json): 16 services / 14 free tier / 1 no key / 14 official MCP. Accept payments, move money, access bank data and query stock, FX and crypto prices.
- [E-commerce](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/ecommerce.json): 8 services / 8 free tier / 0 no key / 6 official MCP. Manage stores, products, orders and marketplace listings.
- [Automation and integration platforms](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/automation-integration.json): 15 services / 12 free tier / 0 no key / 15 official MCP. One connection that exposes thousands of third-party app actions to an agent, with managed auth.
- [Model APIs and inference](https://raw.githubusercontent.com/eater2/ai_agents_api_library/main/catalog/model-inference.json): 18 services / 9 free tier / 0 no key / 7 official MCP. LLM, embedding and multimodal model endpoints an agent can delegate sub-tasks to.

## Optional

- [README](https://github.com/eater2/ai_agents_api_library#readme)
- [HTML version](https://eater2.github.io/ai_agents_api_library/)
- [Official MCP Registry](https://registry.modelcontextprotocol.io/v0/servers): The official Model Context Protocol registry; agents can call its REST API (with OpenAPI spec) to search and list available MCP servers.
- [public-apis/public-apis](https://raw.githubusercontent.com/public-apis/public-apis/master/README.md): A community-curated GitHub list of 1,400+ free public APIs by category, fetchable as raw Markdown for programmatic parsing.
- [APIs.guru OpenAPI Directory](https://api.apis.guru/v2/list.json): An open, unauthenticated directory of 2,500+ public OpenAPI specs; agents fetch list.json to discover APIs and their OpenAPI definitions.
- [punkpeye/awesome-mcp-servers](https://raw.githubusercontent.com/punkpeye/awesome-mcp-servers/main/README.md): A widely-starred curated GitHub list of MCP servers by category, fetchable as raw Markdown for automated discovery.
- [Smithery](https://api.smithery.ai/servers): An MCP server registry and hosting gateway (6,000+ servers) whose REST API lets agents search servers and obtain connection URLs.
- [Glama MCP Directory](https://glama.ai/api/mcp/v1/servers): A large scanned and scored registry of 90,000+ open-source MCP servers, queryable via a bearer-authenticated REST API.
- [mcp.so](https://mcp.so): A community MCP server marketplace and browsable directory site, but it has no publicly documented JSON/REST API for programmatic access.
- [PulseMCP](https://api.pulsemcp.com/v0.1/servers): A directory of 21,800+ MCP servers updated daily; its Sub-Registry REST API (API key required) lets agents list and search servers.
- [llms.txt convention / directory](https://directory.llmstxt.cloud): An open convention for a root-level llms.txt file giving agents a concise Markdown map of a site's key content; a community directory indexes published files, though it exposes no JSON API.
- [Postman API Network](https://www.postman.com/explore): The largest network of public API collections shared on Postman, browsable on the web but with no documented public API for programmatic discovery.
- [RapidAPI Hub (Nokia API Hub)](https://rapidapi.com): A large third-party API marketplace with growing MCP support for agents, but with no public, unauthenticated catalog endpoint for bulk discovery.