Repolode
swirlai

swirlai/swirl-search

Open-source federated AI search and RAG that queries your apps live and answers with sources, without copying data into a vector database.

Open-source alternative to:Glean

Stars
3,049
Forks
292
Open issues
4
Last commit
10 days ago

What is it?

SWIRL is a self-hosted search layer that sends a query to your connected applications at the moment it is asked, collects the results, re-ranks them, and can generate an answer with clickable citations using an LLM of your choice. Because it queries sources in place, there is no vector database to build, no ETL pipeline to maintain, and no second copy of your data to secure.

The Community edition is Apache 2.0 licensed and runs via Docker Compose. It ships with connectors for public sources such as Arxiv, Europe PMC and Google News, and offers Microsoft 365 integration with OAuth2. Re-ranking uses cosine vector similarity with spaCy and NLTK, and results can be stored in SQLite or Postgres for analytics. Queries can run synchronously or asynchronously over a REST API.

Connectors and result mixers are implemented as extensible Python objects, and a pipelined Processor architecture lets you transform queries, responses and results.

Key features

  • 01Queries connected sources live, so data stays where it is.
  • 02Includes 100+ connectors, with Microsoft 365 integration via OAuth2.
  • 03Generates answers with citations using your own OpenAI API key.
  • 04Re-ranks results with cosine similarity, plus duplicate detection.
  • 05Exposes a REST API for synchronous or asynchronous federated queries.
  • 06Connectors and mixers are extensible Python objects.

Use cases

  • Searching a company knowledge base across SharePoint, Confluence and Drive with source links.
  • Building a support assistant that drafts answers grounded in docs and tickets.
  • Running a developer assistant that searches GitHub, Jira and documentation together.

Quick start

  1. 1

    Install and start the Docker app

  2. 2

    Download the compose file

    curl https://raw.githubusercontent.com/swirlai/swirl-search/main/docker-compose.yaml -o docker-compose.yaml
  3. 3

    Pull the images and start SWIRL

    docker-compose pull && docker-compose up
  4. 4

    Open http://localhost:8000/galaxy and log in with admin / password

Things to know

  • The Docker quick start does not retain data or configuration when shut down.
  • Real-time RAG answers require your own OpenAI API key.
  • Community edition lacks the Enterprise three-pass reranker, canonical answers, MCP server and managed support.

Topics

#ai-search#rag#federated-search#llm#python#django#search-engine#enterprise-search#bigquery#federated-query

FAQ

What is swirl-search?
Open-source federated AI search and RAG that queries your apps live and answers with sources, without copying data into a vector database. SWIRL is a self-hosted search layer that sends a query to your connected applications at the moment it is asked, collects the results, re-ranks them, and can generate an answer with clickable citations using an LLM of your choice. Because it queries sources in place, there is no vector database to build, no ETL pipeline to maintain, and no second copy of your data to secure.
Is swirl-search free to use? What license does it have?
swirl-search is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is swirl-search still maintained?
Actively maintained. The last commit to swirl-search was on Oct 27, 2026.
What can I use swirl-search for?
Searching a company knowledge base across SharePoint, Confluence and Drive with source links. Building a support assistant that drafts answers grounded in docs and tickets. Running a developer assistant that searches GitHub, Jira and documentation together.
How do I install swirl-search?
Install and start the Docker app Download the compose file: curl https://raw.githubusercontent.com/swirlai/swirl-search/main/docker-compose.yaml -o docker-compose.yaml Pull the images and start SWIRL: docker-compose pull && docker-compose up Open http://localhost:8000/galaxy and log in with admin / password
Where can I get swirl-search?
The source code and installation instructions are on GitHub: https://github.com/swirlai/swirl-search

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects