How to use chromem-go with Ollama for local embeddings?

Pass `chromem.NewEmbeddingFuncOllama(...)` as the embedding function when creating a collection, as shown in the examples. This allows using locally run models without external API calls.

chromem-go vs Qdrant for Go projects?

chromem-go is embeddable and dependency-free, ideal for lightweight integration in Go apps, while Qdrant is a full client-server database in Rust better for large-scale, distributed deployments. Choose based on scale and architecture needs.

Does chromem-go support persistent storage?

Yes, it offers optional immediate persistence to disk using gob encoding with gzip compression, and backup methods for exporting/importing to single files or blob storage like S3.

What are the performance benchmarks for chromem-go?

Benchmarks show query times of ~0.3 ms for 1,000 documents and ~40 ms for 100,000 documents on a 2020 Intel CPU, with minimal memory allocations, as detailed in the README.

How to implement RAG with chromem-go in Go?

Store documents with embeddings, query for similar content using the Query method, and feed results to an LLM. Example code in the repository demonstrates this workflow with providers like OpenAI or Ollama.

Is chromem-go production ready?

It's in beta with potential breaking changes before v1.0.0, so use with caution in production. However, it's stable for common use cases like prototyping and lightweight deployments.

chromem-go

MPL-2.0Gov0.7.0

An embeddable vector database for Go with a Chroma-like interface, enabling RAG and semantic search without external dependencies.

GitHub

What is chromem-go?

chromem-go is an embeddable vector database library for Go that provides a Chroma-like interface for storing and querying document embeddings. It enables developers to add retrieval augmented generation (RAG), semantic search, and other embeddings-based features directly into their Go applications without requiring a separate database server. The library focuses on simplicity, performance, and zero third-party dependencies, making it ideal for integrating AI-powered search and recommendation capabilities.

Target Audience

Go developers building applications that require local vector database functionality, such as those implementing RAG systems, semantic search, recommendation engines, or classification tasks without relying on external services.

Value Proposition

Developers choose chromem-go for its embeddability, eliminating the complexity of managing a separate vector database server, and its zero-dependency design that ensures lightweight integration. Its Chroma-like API offers familiarity, while performance optimizations and support for multiple embedding providers make it a versatile choice for production use cases.

Overview

Embeddable vector database for Go with Chroma-like interface and zero third-party dependencies. In-memory with optional persistence.

Use Cases

Best For

Related Projects

Community-curated · Updated weekly · 100% open source

Found a gem we're missing?

Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.

Submit a project Star on GitHub

933 stars65 forks0 contributors

Adding retrieval augmented generation (RAG) to Go applications without external dependencies
Building semantic search functionality for text or code within a Go service
Implementing recommendation systems that require efficient similarity search
Creating classification or clustering tools using embeddings in Go
Developing AI-powered features like question answering with up-to-date knowledge bases
Prototyping or deploying lightweight vector search capabilities in resource-constrained environments

Not Ideal For

Applications needing to handle millions of documents efficiently, as exhaustive search scales poorly with size
Projects requiring complex query filters with AND/OR operators, which are not yet implemented
Distributed systems or multi-client setups that benefit from a client-server database architecture
Use cases demanding immediate support for non-text data types like images or videos

Pros & Cons

Pros

Embeddable Design

Runs entirely within your Go application, eliminating the need for a separate database server, similar to SQLite for vector data, which simplifies deployment and maintenance.

Zero Dependencies

No external libraries are required, ensuring lightweight integration without version conflicts, as highlighted in the README's focus on simplicity.

Multithreaded Performance

Leverages Go's concurrency for fast document addition and querying, with benchmarks showing queries on 100,000 documents in ~40 ms on mid-range hardware.

Flexible Embedding Providers

Supports multiple hosted and local embedding services like OpenAI, Azure OpenAI, and Ollama, with a custom interface for easy integration of other models.

Optional Persistence

Offers in-memory operation with optional disk persistence using gob encoding and gzip compression, allowing for data durability without external storage.

Cons

Beta Stability Risks

The project is in beta and may introduce breaking changes before v1.0.0, as warned in the README, making it risky for production use without careful versioning.

Limited Scalability

Relies on exhaustive nearest neighbor search without approximate indices, which can be slow for very large datasets compared to databases with HNSW or IVFFlat support.

Basic Filtering

Only supports simple document and metadata filters like contains and exact matches, lacking advanced operators such as AND/OR, which are still on the roadmap.

Open Source Alternative To

chromem-go is an open-source alternative to the following products:

Pinecone

Pinecone is a vector database service designed for machine learning applications, enabling efficient storage and retrieval of high-dimensional vector embeddings.

C

Chroma

Chroma is an open-source AI-native embedding database and vector store designed for building applications with large language models, enabling semantic search and retrieval-augmented generation.

Q

Qdrant

Qdrant is a vector similarity search engine and vector database designed for machine learning applications, enabling efficient storage and retrieval of high-dimensional vectors.

Weaviate is an open-source vector database that enables semantic search through machine learning models, storing data objects and vectors for similarity-based retrieval.

Frequently Asked Questions

Home

Go

Ollama

Get up and running with Kimi-K2.5, GLM-5, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Stars170,585

Forks15,946

Last commit19 hours ago

LocalAI

LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.

Stars46,013

Forks4,047

Last commit23 hours ago

langchaingo

LangChain for Go, the easiest way to write LLM-based programs in Go

Stars9,179

Forks1,093

Last commit3 months ago

#retrieval-augmented-generation