# Wikimedia Enterprise > The official commercial API for Wikipedia and other Wikimedia projects, from the > Wikimedia Foundation. Structured, machine-readable content across 920+ datasets and > 350+ languages, built for LLM grounding, search, and knowledge graphs. Content is > licensed CC-BY-SA-4.0 and every response carries machine-readable license information. Three products share one access token: Snapshot (bulk downloads), On-demand (single article lookups), and Realtime (streaming updates). Authentication is a JWT bearer token from auth.enterprise.wikimedia.com; data comes from api.enterprise.wikimedia.com. A free account needs no credit card. ## MCP server This documentation as an MCP server: search, fetch, list_endpoints and get_endpoint over the guides and the OpenAPI definition, read-only, no account needed. Streamable HTTP, MCP revision 2026-07-28 with earlier clients served too. - [MCP endpoint](https://enterprise.wikimedia.com/mcp): add it to Claude Code, Cursor, VS Code or ChatGPT as a remote server - [Setup guide](https://enterprise.wikimedia.com/docs/agents.md): client configuration and what each tool returns ## API reference Start here to see exactly what the API returns. The OpenAPI definition is the source of truth for every request and response shape; the reference site renders it with a request playground. - [OpenAPI 3.1 definition (YAML)](https://enterprise.wikimedia.com/docs/api/wme-api.yaml): the machine-readable contract - [API reference](https://enterprise.wikimedia.com/docs/api/): the same definition, rendered and browsable - [Go SDK](https://github.com/wikimedia-enterprise/wme-sdk-go): official client library - [Python SDK](https://github.com/wikimedia-enterprise/wme-sdk-python): official client library ## Documentation Every guide below links to its markdown source. Drop the .md from any of these URLs for the HTML page, and append .md to any /docs/ URL to get the markdown, or request the page URL with Accept: text/markdown. Listed in reading order, not alphabetically. - [Getting Started](https://enterprise.wikimedia.com/docs/index.md): Sign up for a free account, get your JWT tokens, and make your first calls to the Wikimedia Enterprise Snapshot, On-demand, and Realtime APIs. - [Authentication](https://enterprise.wikimedia.com/docs/authentication.md): Use credentials and tokens to manage your Wikimedia Enterprise account programmatically. - [Metadata](https://enterprise.wikimedia.com/docs/metadata.md): Reference endpoints for project codes, languages, projects, and namespaces used across the Wikimedia Enterprise API to build accurate queries. - [Snapshot API](https://enterprise.wikimedia.com/docs/snapshot.md): Download every article in a Wikimedia project as compressed NDJSON files. - [On-demand API](https://enterprise.wikimedia.com/docs/on-demand.md): REST API returning the current version of any article from all supported Wikimedia projects by name, with optional metadata filtering and field selection. - [Realtime API](https://enterprise.wikimedia.com/docs/realtime.md): Stream real-time Wikimedia article and Wikidata updates via SSE or NDJSON, or download hourly batch files of changes across all supported projects, including Wikidata. - [Best Practices](https://enterprise.wikimedia.com/docs/best-practices.md): Practical guidance on authentication, fields, filters, and efficient use of the Snapshot, On-demand, and Realtime Wikimedia Enterprise APIs. - [Identifiers](https://enterprise.wikimedia.com/docs/identifiers.md): The format each Wikimedia Enterprise API resource expects for its identifier - snapshot bundles, chunks, hourly batches, projects, languages, namespaces - and how to look up valid values. - [Fields, Filters, and Limit](https://enterprise.wikimedia.com/docs/filtering.md): Use fields, filters, and limit POST parameters to customize what the Wikimedia Enterprise API returns - scope by project, language, namespace, or field. - [Status Codes](https://enterprise.wikimedia.com/docs/status-codes.md): Overview of HTTP response status codes returned by the Wikimedia Enterprise APIs, including success, redirection, client error, and server error codes. - [Data Primer](https://enterprise.wikimedia.com/docs/data-primer.md): What are the differences between Wikimedia, Wikipedia, and MediaWiki? Who writes and edits all these articles? What API should I use to access article content? - [Data Model & Dictionary](https://enterprise.wikimedia.com/docs/data-dictionary.md): How a Wikimedia Enterprise API response is shaped: the article object at its center, the parsed and Wikidata variants beside it, and where every field is defined. - [Tuning Credibility Signals](https://enterprise.wikimedia.com/docs/credibility-signals.md): How to put credibility signals to work in your pipeline: recommended thresholds, an evaluation strategy, and field-by-field reference tables. - [AI Agents and MCP](https://enterprise.wikimedia.com/docs/agents.md): How coding agents and AI assistants read this documentation: the llms.txt index, markdown versions of every guide, and the docs MCP server with its four tools. ## Products and datasets What the API covers, by product and by Wikimedia project. Read these to judge fit before reading the endpoint docs. - [API Products](https://enterprise.wikimedia.com/api/): Our APIs include 130M+ pages across 920+ Wikimedia projects and 350+ languages - [Structured Contents](https://enterprise.wikimedia.com/api/structured-contents/): Wikipedia uses wikitext, a markup language designed for formatting page content. - [Credibility Signals](https://enterprise.wikimedia.com/api/credibility-signals/): Credibility signals are data points in every Wikimedia Enterprise API response that help you judge whether a Wikipedia edit is safe to use and significant enough to act on. - [Project Data](https://enterprise.wikimedia.com/project-data/): Wikipedia and Wikidata are what most people come for: the encyclopedia, and the knowledge graph that holds the structured facts behind it. - [Wikipedia](https://enterprise.wikimedia.com/project-data/wikipedia-api/): Wikipedia is the largest single source of human-written knowledge on the internet: over 68 million articles across 348 language editions, edited about five times a second, and governed by verifiability and neutral point of view. - [Wikidata](https://enterprise.wikimedia.com/project-data/wikidata-api/): The Wikidata API gives you programmatic access to Wikidata, the Wikimedia project that stores the structured data behind Wikipedia, Wikivoyage, Wiktionary, Wikisource, and every other Wikimedia project. - [Wikisource](https://enterprise.wikimedia.com/project-data/wikisource-api/): Wikisource is the Wikimedia sister project dedicated to collecting and preserving previously published source texts, spanning constitutions, treaties, seminal scientific papers, poetry, drama, religious texts, and the great works of world literature. - [Wiktionary](https://enterprise.wikimedia.com/project-data/wiktionary-api/): Wiktionary is the free multilingual dictionary, and it is the largest one ever assembled: over 49 million entries across 174 language editions, more than nine million of them in the English edition alone. - [Wikiquote](https://enterprise.wikimedia.com/project-data/wikiquote-api/): Wikiquote is the free compendium of sourced quotations: over 380,000 pages of quotes from notable people and from creative works, gathered across 77 language editions and held to the same sourcing and notability standards as Wikipedia. - [Wikibooks](https://enterprise.wikimedia.com/project-data/wikibooks-api/): Wikibooks is a Wikimedia project dedicated to providing free, open-content textbooks, manuals, and other instructional texts. - [Wikiversity](https://enterprise.wikimedia.com/project-data/wikiversity-api/): Wikiversity is the Wikimedia project for learning: course modules, lesson plans, study guides, learning activities, and original research, written by teachers and students for every level of education from pre-school to university. - [Wikivoyage](https://enterprise.wikimedia.com/project-data/wikivoyage-api/): Wikivoyage is a sister project of Wikipedia dedicated to travel, containing over 142,000 articles. ## Pricing Free and paid accounts, with the plan facts in markdown so they can be quoted verbatim. Paid egress pricing is bespoke and is not published. - [Pricing (Markdown)](https://enterprise.wikimedia.com/pricing.md) - [Pricing (HTML)](https://enterprise.wikimedia.com/pricing/) ## Company and terms Who runs this and under what agreement. The legal pages are HTML only; they are plain prose and read fine as is. - [About Wikimedia Enterprise](https://enterprise.wikimedia.com/about/): Wikimedia Enterprise is a product of the Wikimedia Foundation, built to make Wikimedia project data, created by volunteers around the world, accessible at scale through enterprise-grade APIs. - [Contact](https://enterprise.wikimedia.com/contact/): reach the team about a paid plan or a technical question - [API License Agreement](https://enterprise.wikimedia.com/terms/): Please read this API License Agreement ("Agreement") carefully before using the Enterprise APIs offered by Wikimedia, LLC ("Wikimedia"). - [Privacy Policy](https://enterprise.wikimedia.com/privacy/): We at Wikimedia LLC ("Wikimedia") know you care about how your personal information is used and shared in relation to enterprise.wikimedia.com or our application programming interfaces (collectively, "Wikimedia Enterprise"), and we take your privacy seriously. ## Optional Background and history. Useful for tracking what changed and how the data is produced, not needed to integrate. Both feeds are the cheapest way to watch for changes. - [API changelog](https://enterprise.wikimedia.com/docs/changelog/): dated notes on endpoint and payload changes - [API changelog feed (RSS)](https://enterprise.wikimedia.com/docs/changelog/rss.xml) - [Blog](https://enterprise.wikimedia.com/blog/): product announcements and technical write-ups - [Blog feed (RSS)](https://enterprise.wikimedia.com/blog/rss.xml)