Using Local Ollama

Docxa allows you to run your entire documentation pipeline locally using Ollama. This is ideal for sensitive codebases or when working offline.

Setup Ollama

  1. Install Ollama: Download from ollama.com.
  2. Pull a Model: We recommend llama3.1 or mistral.
    ollama pull llama3.1

Configure Docxa

Set your environment variables to point to your local instance:

export DOCXA_PROVIDER=ollama
export DOCXA_MODEL=llama3.1

Advanced: Custom Endpoint

If Ollama is running on a different port or machine, use DOCXA_OLLAMA_URL:

export DOCXA_OLLAMA_URL=http://my-server:11434/v1

[!TIP] Docxa uses the OpenAI-compatible API provided by Ollama, ensuring seamless switching between local and cloud providers.

results matching ""

    No results matching ""