Providers

OpenWalrus supports multiple LLM providers through a unified Model trait. All providers are API-based — configure an endpoint, point the daemon at it, and go.

Two API standards

All providers use one of two wire formats, selected by the api_standard field in config:

If api_standard is omitted, OpenWalrus defaults to openai. If the base_url contains "anthropic", the Anthropic standard is auto-detected.

Provider configuration

Each provider is a [model.<name>] section in walrus.toml:

[model.deepseek]
model = "deepseek-chat"
api_key = "${DEEPSEEK_API_KEY}"

Configure multiple providers and switch between them:

[model.gpt-4o]
model = "gpt-4o"
api_key = "${OPENAI_API_KEY}"

[model.claude]
model = "claude-sonnet-4-20250514"
api_key = "${ANTHROPIC_API_KEY}"
api_standard = "anthropic"

Any model name is valid — the api_standard field (not the model name) determines which API protocol to use.

Selecting the active model

Set the default model in the [walrus] section:

[walrus]
model = "claude"

Override per agent in [agents.*]:

[agents.researcher]
model = "deepseek"

Provider manager

The ProviderManager holds all configured providers and routes requests by model name. It supports hot-reload — update the config and the active provider changes without restarting the daemon.

What's next

• Remote providers — OpenAI, Claude, DeepSeek, and more
• Configuration — full config setup