https://github.com/maragudk/gai-starter-kit
Get started with LLMs, FTS and vector search, RAG, and more, in Go!
https://github.com/maragudk/gai-starter-kit
ai evals fts go llm rag sqlite vector-search
Last synced: 23 days ago
JSON representation
Get started with LLMs, FTS and vector search, RAG, and more, in Go!
- Host: GitHub
- URL: https://github.com/maragudk/gai-starter-kit
- Owner: maragudk
- License: mit
- Created: 2025-03-05T11:58:46.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-13T09:44:27.000Z (about 1 year ago)
- Last Synced: 2025-03-13T10:34:42.356Z (about 1 year ago)
- Topics: ai, evals, fts, go, llm, rag, sqlite, vector-search
- Language: Go
- Homepage:
- Size: 286 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Get started with LLMs, FTS and vector search, RAG, and more, in Go!
[](https://pkg.go.dev/github.com/maragudk/gai-starter-kit)
[](https://github.com/maragudk/gai-starter-kit/actions/workflows/ci.yml)
[](https://github.com/maragudk/gai-starter-kit/actions/workflows/cd.yml)
Made with ✨sparkles✨ by [maragu](https://www.maragu.dev/).
Does your company depend on this project? [Contact me at markus@maragu.dk](mailto:markus@maragu.dk?Subject=Supporting%20your%20project) to discuss options for a one-time or recurring invoice to ensure its continued thriving.
## Overview
This is a template application for developers interested in building Go web applications with:
- Large Language Models (LLMs) and foundation models integration
- Document search capabilities using both full-text search (BM25) and vector search (embeddings)
- A flexible architecture supporting RAG (Retrieval Augmented Generation) and tool use
Key features:
- Local database (SQLite) for document storage and retrieval
- Local LLM support (Llama 3) for text generation
- Local embeddings model (mxbai-embed-large-v1) for vector generation
- Document CRUD endpoints with automatic chunking
- Simple and extensible Go architecture
## Roadmap
- [x] Local SQLite database with full-text search (FTS5)
- [x] Local LLM integration (Llama 3)
- [x] Local embeddings model (mxbai-embed-large-v1)
- [x] Document CRUD API with automatic chunking
- [x] Vector search implementation
- [ ] Prompt endpoint with LLM tool use capabilities
- [ ] RAG implementation for improved LLM responses
- [ ] Advanced chunking strategies
- [ ] Multi-model support
## Evals
