Cohere Toolkit

Toolkit is a collection of prebuilt components enabling users to quickly build and deploy RAG applications.

Try Now:

Try the default Toolkit application yourself by deploying it in a container locally. Either with docker run:

docker run -e COHERE_API_KEY='>>YOUR_API_KEY<<' -p 8000:8000 -p 4000:4000 ghcr.io/cohere-ai/cohere-toolkit:latest

or cloning and running locally:

git clone https://github.com/cohere-ai/cohere-toolkit.git
make first-run

Go to localhost:4000 in your browser and start chatting with the model.

Interfaces - these can be any frontend, application, bot or integration. You can customize any type of interface for your use case. By default included is:
- Cohere's Web UI at src/interfaces/coral_web - A web app built in Next.js. Includes a simple SQL database out of the box to store conversation history in the app.
Backend API - src/backend this follows a similar structure to the Cohere Chat API but also include customizable elements:
- Model - you can customize with which provider you access Cohere's Command models. By default included in the toolkit is Cohere's Platform, Sagemaker, Azure, hugging face, local models. More details here.
- Retrieval- you can customize tools and data sources that the application is run with. By default, we have configured a Langchain data retriever to test RAG on Wikipedia and your own uploaded documents. It is possible to add any tool including any tools or retrievers from LangChain or LlamaIndex. You can also use a connector you have created.
Service Deployment Guides - we also include guides for how to deploy the toolkit services in production including with AWS, GCP and Azure. More details here.

Contributions are what drive an open source community, any contributions made are greatly appreciated. To get started, check out our documentation.