Scry

Scry isn't a regular search engine — it's infrastructure for programmatic querying of the internet. It gives your agent access to over 164 billion queryable rows from Reddit, Hacker News, arXiv, Stack Exchange, Wikipedia, and prediction markets, with millions of new records added daily. Instead of fetching one page at a time, your agent writes a single query that runs across billions of documents. It supports vector-based semantic search, full relational queries, and even graph walks through reply and discussion threads. Connects via MCP server directly to ChatGPT, Claude, Cursor, and Claude Code. Pricing floats with congestion: heavier queries cost more. The free researcher tier is sufficient for independent work.