- Home
- Vector Databases
- Vearch
Vearch
What it is
Vearch is a free, open-source distributed vector database for similarity search in AI applications. It searches embedding vectors and can combine those searches with scalar filtering for metadata conditions. The project describes replication and elastic scale-out, and says it can retrieve vectors among millions of objects in milliseconds. Its Router exposes REST operations for upserting, deleting, searching, and querying data. SDKs are listed for Python, Go, Java, and Rust, with integrations for LangChain, LlamaIndex, LangChainGo, and LangChain4j. Deployment options include Kubernetes through Helm, Docker Compose in standalone or cluster mode, Docker, and source compilation. PartitionServer stores document partitions with Raft-based replication, while Gamma provides vector search based on Faiss. Vearch uses the Apache License, Version 2.0 and supports self-hosted operation. Community help is available through GitHub Issues, an email discussion group, and Slack. The VisualSearch demo additionally requires an image retrieval plugin for detection and feature extraction.
For and against
For
- Combines vector search with scalar filtering
- Lists SDKs for four programming languages
- Supports Kubernetes, Docker, and source deployment
- Provides REST operations through its Router
- Apache License 2.0 open-source project
Against
- VisualSearch requires an additional image retrieval plugin
- Databases cannot be deleted while containing a space
- Repository has no published security policy
The verdict
Vearch offers a broad self-hosted toolkit for vector search, including filtering, replication, SDKs, and deployment choices. Users should account for the VisualSearch dependency and the documented database deletion restriction.
Who it suits
Vearch suits AI teams needing a self-hosted vector database with metadata filtering, client SDKs, and several deployment methods. It is aimed at developers integrating similarity search into applications.
Plans and prices
Vearch
Apache License, Version 2.0 · self-hosted deployment
Read 10 Oct 2026 · github.com
On the record
- API
- The Router exposes REST endpoints for upsert, delete, search, and query operations.github.com · read 10 Oct 2026
- Community support
- The project points users to GitHub Issues, an email discussion group, and a Slack channel for questions and support.github.com · read 4 Oct 2026
- Deployment
- Documented deployment options include Kubernetes with Helm, Docker Compose, Docker, and source compilation.github.com · read 10 Oct 2026
- Framework integrations
- The README lists LangChain, LlamaIndex, Langchaingo, and LangChain4j integrations.github.com · read 10 Oct 2026
- Hybrid search
- It supports vector search together with scalar filtering.github.com · read 10 Oct 2026
- Integrations
- The project lists integrations with LangChain, LlamaIndex, LangChainGo, and LangChain4j.github.com · read 4 Oct 2026
- License
- Vearch is licensed under Apache License, Version 2.0.github.com · read 10 Oct 2026
- Notable limit
- The Vearch 3.5 database guide says a database cannot be deleted while it contains a space.vearch.readthedocs.io · read 10 Oct 2026
- Product
- Vearch is a cloud-native distributed vector database for similarity search of embedding vectors in AI applications.github.com · read 10 Oct 2026
- Scale and reliability
- The project describes replication and elastic scale-out as features.github.com · read 10 Oct 2026
- SDKs
- The README lists Python, Go, Java, and Rust SDKs.github.com · read 10 Oct 2026
- Search performance
- The project says it can retrieve vectors among millions of objects in milliseconds.github.com · read 10 Oct 2026
- Security policy
- GitHub's security page says the repository has no SECURITY.md policy and no published security advisories.github.com · read 10 Oct 2026
- Storage engine
- PartitionServer hosts document partitions with Raft-based replication, and Gamma is the core vector search engine based on Faiss.github.com · read 10 Oct 2026
- Support
- The project directs users to GitHub Issues, a public email discussion list, and Slack for community support.github.com · read 10 Oct 2026
- Supported build environments
- Vearch's installation guide says CentOS, Ubuntu, and Mac OS are supported for compiling the software.vearch.readthedocs.io · read 10 Oct 2026
- Visual search requirement
- The VisualSearch demo requires an image retrieval plugin for object detection and feature extraction.github.com · read 4 Oct 2026
