DEV.co
RAG Frameworks · swirlai

swirl-search

SWIRL is an open-source federated search and RAG platform that queries your existing data sources live—SharePoint, Confluence, Jira, Google Drive, etc.—without copying data into a vector database. It ranks results and optionally generates answers with citations, all via a web UI or REST API.

Source: GitHub — github.com/swirlai/swirl-search
3k
GitHub stars
285
Forks
Python
Primary language
Apache-2.0
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositoryswirlai/swirl-search
Ownerswirlai
Primary languagePython
LicenseApache-2.0 — OSI-approved
Stars3k
Forks285
Open issues2
Latest releasev4.5.0.7 (2026-06-22)
Last updated2026-06-29
Sourcehttps://github.com/swirlai/swirl-search

What swirl-search is

Django-based Python application supporting 100+ connectors for live federated queries across disparate sources. Features cosine similarity re-ranking (spaCy/NLTK), Processor pipeline architecture for transforming queries/responses, SQLite/Postgres result storage, OAuth2 for M365, and extensible Connector/Mixer objects. Deployable via Docker in minutes.

Quickstart

Get the swirl-search source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/swirlai/swirl-search.gitcd swirl-search# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Enterprise Knowledge Base Search

Unified search across SharePoint, Confluence, and Google Drive with source links and permission inheritance. Eliminates separate search silos and respects existing access controls at query time.

Customer Support and Documentation Search

Search knowledge articles, support tickets, and docs simultaneously; generate grounded responses with citations. No ETL pipeline required—queries run live against existing systems.

Developer Tool Integration

Index GitHub repos, Jira issues, and internal documentation in a single search interface. Useful for onboarding, troubleshooting, and cross-team knowledge discovery without data replication.

Implementation considerations

  • Ensure all upstream data sources (SharePoint, Confluence, Jira, etc.) have stable API access and OAuth2/auth credentials ready before deployment.
  • Plan connector configuration and custom Processor logic for query rewriting, response deduplication, and result ranking tuning specific to your content.
  • Validate that live query latency across your sources meets user expectations; implement caching or result post-processing if federation is too slow.
  • Establish governance for API key and credential management (OpenAI, M365 OAuth, connector secrets); use environment variables or a secrets manager.
  • Configure SQLite (development) or Postgres (production) backend for result persistence and analytics; plan data retention and cleanup policies.

When to avoid it — and what to weigh

  • Sub-Second Latency Requirement — Live federated queries introduce latency due to upstream source response times. If real-time SLAs demand <500ms results, vector DB pre-indexing may be mandatory.
  • Offline or Disconnected Operation — SWIRL requires live connectivity to all source systems. Cannot operate if sources are intermittently unavailable or your use case needs offline-first search.
  • Highly Specialized Ranking Needs — Community uses basic cosine similarity ranking. If you need advanced BM25, embeddings fusion, cross-encoder re-ranking, or domain-specific scoring, Enterprise is required.
  • Managed Infrastructure Requirement — Community is self-hosted only. If your org requires vendor-managed SaaS, SOC-2 compliance infrastructure, or fully managed connectors with guaranteed uptime SLAs, only Enterprise qualifies.

License & commercial use

Apache License 2.0 (Apache-2.0). Permissive OSI-approved license allowing commercial use, modification, and redistribution with attribution and liability disclaimer. No copyleft obligations. Full license available in the repository.

Apache 2.0 permits commercial use without vendor approval or licensing fees. You may deploy SWIRL Community in production for internal or customer-facing purposes, modify source code, and redistribute derivatives—provided you retain Apache headers and NOTICE files. No warranty is provided. Commercial support and managed hosting are available via SWIRL Enterprise; Community relies on community support channels.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityLow
DEV.co fitStrong
Assessment confidenceHigh
Security considerations

SWIRL enforces source-level permissions (queries run with user's own credentials against upstream systems). No data is copied or stored centrally, reducing exposure surface. Community version offers no built-in encryption, hallucination warnings, or formal security audit claims. OAuth2 M365 integration available. Plan for secure credential storage (env vars, secrets manager) for API keys and connector tokens. Network traffic between SWIRL and sources should be encrypted (HTTPS). No SOC-2 or formal security certifications in Community; Enterprise offers SOC-2 hosted option.

Alternatives to consider

Elasticsearch with custom connectors

Requires ETL and data movement but offers mature indexing, powerful search syntax, and horizontal scaling. Choose if you can tolerate centralized copies or need sub-second latency.

Pinecone / Weaviate / Milvus (Vector DBs)

Specialized for semantic search and RAG but require data ingestion and vector embedding pipelines. Choose if semantic similarity is your primary requirement and latency is acceptable.

SWIRL Enterprise

Same core federated search but adds three-pass re-ranking, canonical answers, MCP server for agents, managed connectors, and SOC-2 hosting. Choose if you outgrow basic cosine ranking or need vendor-managed infrastructure.

Software development agency

Build on swirl-search with DEV.co software developers

SWIRL's Apache 2.0 license and Docker deployment let you start in minutes. Evaluate on your own data with no vendor lock-in. For production scale with advanced ranking and managed connectors, explore SWIRL Enterprise.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

swirl-search FAQ

Do I have to move my data to use SWIRL?
No. SWIRL queries your sources live and in place. Data stays in SharePoint, Confluence, Jira, etc. No ETL or vector database required.
What LLM does SWIRL use for RAG?
Community supports your own OpenAI key (via OPENAI_API_KEY environment variable). Other LLM providers are not explicitly mentioned in the provided data; see docs for current support.
Can I use SWIRL in production without paying?
Yes, under Apache 2.0. SWIRL Community is free to self-host in production. Commercial support and managed infrastructure are available via Enterprise.
How fast is SWIRL search?
Latency depends on your upstream source response times (live federation). Community uses cosine similarity re-ranking. Enterprise offers three-pass re-ranking for potentially better results but not necessarily faster. Sub-second guaranteed latency is not a strength; consider a vector DB if speed is critical.

Work with a software development agency

Need help beyond evaluating swirl-search? DEV.co is a software development agency offering software development services and web development for teams of every size. Our software developers and web developers build custom software, web applications, APIs, and rag frameworks integrations — and maintain them long-term.

Ready to Build Federated Search Into Your Stack?

SWIRL's Apache 2.0 license and Docker deployment let you start in minutes. Evaluate on your own data with no vendor lock-in. For production scale with advanced ranking and managed connectors, explore SWIRL Enterprise.