ChatLocalAI

LocalAI Setup

LocalAI is a drop-in replacement REST API that’s compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families that are compatible with the ggml format.

To use ChatLocalAI within Flowise, follow the steps below:

1. git clone https://github.com/go-skynet/LocalAI

2. cd LocalAI

3. # copy your models to models/
   cp your-model.bin models/

For example:

Download one of the models from gpt4all.io

# Download gpt4all-j to models/
wget https://gpt4all.io/models/ggml-gpt4all-j.bin -O models/ggml-gpt4all-j

In the /models folder, you should be able to see the downloaded model in there:

Refer here for list of supported models.

1. docker compose up -d --pull always

2. Now API is accessible at localhost:8080

# Test API
curl http://localhost:8080/v1/models
# {"object":"list","data":[{"id":"ggml-gpt4all-j.bin","object":"model"}]}

Flowise Setup

Drag and drop a new ChatLocalAI component to canvas:

Fill in the fields:

• Base Path: The base url from LocalAI such as http://localhost:8080/v1

• Model Name: The model you want to use. Note that it must be inside /models folder of LocalAI directory. For instance: ggml-gpt4all-j.bin

If you are running both Flowise and LocalAI on Docker, you might need to change the base path to http://host.docker.internal:8080/v1. For Linux based systems the default docker gateway should be used since host.docker.internal is not available: http://172.17.0.1:8080/v1

That's it! For more information, refer to LocalAI docs.

Watch how you can use LocalAI on Flowise