Models

models.toml declares providers and model definitions. Each provider’s api_key lives in secrets.toml under a matching [providers.<name>] table — this file only holds base_url and the model catalog.

Each [[models]] block should include:

model: the provider-specific model name sent in API requests.
task: the task this model serves (chat, embedding, image-generation, text-to-speech, speech-to-text, rerank).
provider (optional): the provider key from [providers.*] to use; defaults to whichever provider has default = true.

Each [providers.*] table may set default = true to mark it as that fallback. If several tables set it, the first one (file order) wins.

Example:

[providers.ollama]
default = true
base_url = "http://localhost:11434/v1"

[[models]]
model = "llama3.2:1b"
task = "chat"

Ollama needs some api_key string even though it ignores its value — set it in secrets.toml:

[providers.ollama]
api_key = "ollama"

Notes:

The shipped templates cover local Ollama (llama3.2:1b, qwen2.5:0.5b, qwen3:1.7b, llama3.2:3b) and hosted Fireworks.
To switch which model handles a task, set models.<task> in settings.toml directly to that model’s {model, provider} pair.

Services