WHAT IS IT?
OpenSERP is a self-hosted SERP API and CLI that fetches live search results from six engines (Google, Bing, Yandex, Baidu, DuckDuckGo, Ecosia) without API keys or per-search billing. You run it with Docker or as a Go binary, then hit an HTTP endpoint that returns results in a unified JSON schema, whatever engine sits behind the query.
WHY IS IT INTERESTING?
- Six engines, one schema: responses from every engine are normalized into the same JSON format, so your code stays identical when you switch from Google to Bing.
- Megasearch: a single endpoint queries several engines in parallel, with deduplication and clustering of the results.
- Content extraction: the tool can visit the pages it finds and embed their cleaned-up content as Markdown, ready to feed an LLM or a RAG pipeline.
- Headless with fallback: browser rendering for JavaScript-heavy pages, raw HTTP mode as a backup, HTTP and SOCKS5 proxy support, built-in caching.
- Built for AI agents: MCP server, SDKs, n8n integration and JSON, Markdown, text or NdJSON output formats.
USE CASES
- Give an LLM agent live web access without paying for a commercial search API.
- Track the SEO ranking of your pages across several engines at once.
- Feed a RAG pipeline with fresh content already extracted as Markdown.
- Automate multi-engine competitive monitoring through the HTTP API.
- Script image searches with language, date and file-type filters.
