# Wikexa — Structured Knowledge API for AI Agents

Pre-processed Wikipedia and Wikidata served as clean JSON at the edge. Built so
an agent spends tokens on facts instead of on HTML.

## Why

Fetching https://en.wikipedia.org/wiki/Apple_Inc. costs roughly 15,000 tokens of
HTML to recover about 500 tokens of facts. One lookup() call here returns the
facts directly.

## MCP

    POST https://wikexa.com/mcp

JSON-RPC 2.0. Methods: initialize, tools/list, tools/call, ping.
No API key. No auth. Free.

## Tools

- lookup(entity) — Look up any entity (person, company, place, concept) and get structured facts plus a ~200-token summary.
- article(title, sections, max_chars) — Full Wikipedia article as clean structured JSON — sections as an array, infobox as key/value facts, no HTML or wikitext.
- search(query, limit) — Find entities by fuzzy text match when you do not know the exact title.
- recent(topic, hours, limit) — Articles changed recently, optionally filtered by topic.

## REST (same handlers as MCP)

    GET https://wikexa.com/v1/lookup?entity=Tim%20Cook
    GET https://wikexa.com/v1/article?title=Apple%20Inc.&sections=History
    GET https://wikexa.com/v1/search?query=quantum%20computing&limit=5
    GET https://wikexa.com/v1/recent?topic=election
    GET https://wikexa.com/v1/stats

## Token control

article() takes `sections` and `max_chars`. Ask for the section you need
rather than the whole article.

## Licensing

Wikipedia text: CC BY-SA 4.0 — attribution required, keep `source_url`.
Wikidata facts: CC0.

## Related

Market data for agents (paid, x402 USDC): https://cabrini.ai/mcp