ContextWire is a free search API for AI developers, positioned as "the context API that actually answers questions." It unifies multi-source research engines into a single interface, helping AI applications answer questions accurately based on real sources.
In terms of use, ContextWire mainly provides AI developers and AI applications with web search, web page reading, and answer extraction capabilities, making it suitable for scenarios requiring factual question answering, research, and source citations.
Main capabilities include:
- Multi-engine search: Connects to Google and Bing through SearXNG, supports automatic retries and result deduplication, and provides 22 search configurations covering categories such as web, news, academic, code, and social.
- Deep page reading: When summaries are not enough, fetches and reads the full web page, extracts relevant paragraphs and supports tables, and outputs clean Markdown or plain text.
- LLM answer extraction: Uses the Trinity model to extract precise answers from context, runs on the OpenRouter free tier, is zero-cost, and requires no API key management.
- Academic search: Integrates CrossRef, arXiv, Semantic Scholar, PubMed, and OpenAlex, and can retrieve DOIs, citations, abstracts, and metadata.
- Game and media wikis: Directly accesses the Fandom and wiki.gg APIs, covering Minecraft, Dota 2, Dragon Age, Rick & Morty, and more.
- In-depth Wikipedia parsing: Directly calls the Wikipedia API, supporting table extraction, infobox parsing, and Wikidata entity resolution.
How to use: The API provides a single endpoint and returns JSON responses. It can be added as an MCP server, and with one command you can use five tools in Claude Code or Cursor: ask, search, extract, research, and batch search. An online API tester is also provided, allowing you to try each endpoint with real data and copy code snippets.
Billing and quota: The free plan offers 1,000 queries per month; bringing your own OpenRouter key (BYOK) consumes 1 credit per query, while not using your own key consumes 2 credits per query. There is also a Pro plan with 3,500 queries/month, including batch parallel search, deep research mode, priorityrouting; and a higher-tier plan with 15,000 queries/month, including custom search configurations, dedicated support, and usage analytics. The official statement says you can get a free API key without a credit card. Unlike Tavily or the Perplexity API, ContextWire does not lock users into using its LLM, and you can connect your own provider.
Applicable scenarios: Adding real-time factual question answering and source citations to AI applications; building answer engines, fact-checking tools, and research report generators; conducting real-time research in AI programming environments.
Limitations: The materials do not specify exact pricing, response latency, terms of service, or a complete list of data coverage; the free quota and features are subject to the official page.
