#10 · Primary category: Customer Support

simba

customer-service evals knowledge-base llm rag

OpenSource Production ready Customer service with built in Evals and monitoring

Project last updated:06/18/26

GitHub Stars

1.5K

Forks

114

Contributors

8

License

Apache-2.0

Why we included this project

Simba is an open-source customer service assistant for teams that want to run their own AI support channel instead of renting a black-box vendor product. It pairs a customizable retrieval pipeline, where you can swap vector stores, embedding models, LLMs, and rerankers, with a drop-in npm chat widget, so a working assistant can be embedded on a website in a single integration step. It goes beyond a plain RAG demo thanks to the built-in evaluation layer: retrieval precision and recall, generation faithfulness, and latency are tracked out of the box, giving you concrete signals to iterate on rather than guessing how well the assistant answers. A dashboard and conversation analytics let you monitor real usage, and because it self-hosts, you keep control over your data and pipeline components. Small product and support teams that want a production baseline they can own and adapt will find it a practical fit.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category