nothumansearch
An open-source search engine for AI agents that helps developers find websites and interfaces that are easier for agent and RAG workflows to consume programmatically.
Tool overview
Based on the available evidence, nothumansearch is best classified as search infrastructure for AI agents, not a general-purpose web search engine, a chat-answer product, or a full enterprise knowledge platform. The GitHub repo description explicitly says it ranks sites using agent-readiness signals such as llms.txt, OpenAPI, MCP, and ai-plugin support, and provides an MCP server, REST API, full-text search, plus 8,000+ indexed sites. That supports the positioning, but it does not yet prove production-grade retrieval quality at scale.
In practice, its value seems to be helping agents discover websites and interfaces that are easier to consume programmatically. If you are building AI agents, tool-use workflows, crawler-assisted retrieval, or machine-oriented navigation, this may be more relevant than traditional search. A better analogy is a developer-facing search index with agent-ready ranking logic, not a human answer engine like Perplexity and not just a vector database.
On cost and adoption friction, the evidence only supports that it is open source and exposes MCP/REST interfaces.
This tool does not have related social references to display yet.