Built because I needed one search API that wasn't Google-only.

OpenSERP started in 2023 as a tool for my own projects. It now returns structured results from six engines, remains MIT-licensed, and can run on your machine with one Docker command.

How it got here

OpenSERP grew from a personal tool into a project teams ship on. Two things never changed: it stays open source, and the response format stays stable.

  1. It started as a scraper I needed myself

    Every SERP API I tried was Google-only, paid per query, or both. I wanted one tool that returned structured results from several engines and could run on my own machine. So I built it and put it on GitHub under MIT.

  2. People started trusting the response shape

    Image search, more engines, and page extraction landed because real users asked for them. The JSON shape stayed stable enough that teams could build on top of it without rewrites.

  3. Cloud, for teams that don't want to run servers

    OpenSERP Cloud is the managed path - same data model, hosted keys, a dashboard, and billing. The open-source engine stays free and self-hostable. You can start on one and move to the other without changing your code.

Use or inspect the project

Packages and their source live together here, so each project is listed once.

Read the docs →

Run it

Use an SDK

Connect a workflow