IntelliBar

Documentation

Using custom models

You can add custom models that aren't available in the app by default.

How to add custom models

Custom models are .yaml files in ~/.intellibar/models.

Here's an example of a model file:

# ~/.intellibar/models/deepseek-r1-distill-qwen-32b.yaml
url: https://openrouter.ai/api/v1
id: deepseek/deepseek-r1-distill-qwen-32b
name: DeepSeek R1 32B
apiKey: sk-or-v1-xxxx

Supported fields

  • url — The URL of the endpoint. Can be both remote and localhost.
  • id — The model id/name that will be passed to the endpoint.
  • apiKey — The API key that will be passed to the endpoint. Defaults to n/a. Optional.
  • apiType — The type of API. Possible values: openai, anthropic, google. Optional. Defaults to openai.
  • name — A friendly name that will be displayed inside the app. Defaults to using id. Optional.

Common use cases

  • Models on OpenRouter
  • Models on Azure / AWS / HuggingFace
  • Local models with llama.cpp
  • Local models with Ollama running on a custom port

Examples

url: https://openrouter.ai/api/v1
id: deepseek/deepseek-r1-distill-qwen-32b
name: DeepSeek R1 32B
apiKey: sk-or-v1-xxxx

url: https://api.anthropic.com
id: claude-3-7-sonnet-latest
name: Claude 3.7 Sonnet
apiType: anthropic
apiKey: sk-ant-...

url: http://localhost:8080/
id: gemma3:4b
name: Gemma 3 4B