how to set .env file if i want to use local OpenAI-compatible endpoints

[talentam]

Git provider (optional)

Gitlab

System Info (optional)

No response

Issues details

how to set .env file if i want to use local OpenAI-compatible endpoints

[qodo-merge-for-open-source]

💡 Auto-generated documentation-based answer:

To use local OpenAI-compatible endpoints with PR-Agent, you need to configure the API base URL and model settings. Here's how:

For Local Models (e.g., Ollama):

1. Set the model configuration to use your local endpoint:

   [config]
   model = "ollama/qwen2.5-coder:32b"
   fallback_models = ["ollama/qwen2.5-coder:32b"]
   custom_model_max_tokens = 128000
   
   [ollama]
   api_base = "http://localhost:11434"  # or your custom port

2. For environment variables (if not using a config file):

   CONFIG__MODEL="ollama/qwen2.5-coder:32b"
   OLLAMA__API_BASE="http://localhost:11434"

For OpenAI-like API endpoints:

If you're using a different OpenAI-compatible endpoint:

[openai]
api_base = "http://your-local-endpoint:port/v1"
api_key = "your-key-if-needed"

Or as environment variables:

OPENAI__API_BASE="http://your-local-endpoint:port/v1"
OPENAI__KEY="your-key-if-needed"

Important Notes:

• For GitHub Actions, you'll need a self-hosted runner since hosted runners cannot access localhost services
• Local open-source models are suitable for experimentation but not recommended for production-level code analysis
• Make sure to set custom_model_max_tokens appropriately for your model
• Context window size may need adjustment (e.g., for Ollama, use OLLAMA_CONTEXT_LENGTH environment variable)

For more details on model configuration, see the "Changing a model in PR-Agent" documentation.

Relevant Sources:

• https://qodo-merge-docs.qodo.ai//usage-guide/changing_a_model#ollama
• https://qodo-merge-docs.qodo.ai//usage-guide/changing_a_model#openai-like-api
• https://qodo-merge-docs.qodo.ai//installation/github#using-local-models-ollama