Provider and model routing

Configure Kimi Code, OpenRouter, Ollama Cloud, Codex OAuth, Anthropic, OpenAI, local Ollama, and fallback model routes in Libre Claw.

OpenRouter

Broad model choice, provider-reported usage, automatic context metadata, and fallback routing.

bash OpenRouter setup

libre-claw auth set-key openrouter
/model openrouter:z-ai/glm-5.2 --global

Kimi Code

Kimi membership access through the official coding endpoint, with native reasoning continuity, visual input, tool calls, and published context limits.

bash Kimi Code setup

libre-claw auth set-key moonshot
/model moonshot:k3 --global

Ollama Cloud

Ollama-style cloud model names without running local GPUs.

bash Ollama Cloud setup

libre-claw auth set-key ollama
/model ollama:glm-5.2:cloud --global

Codex OAuth

Use the Codex CLI login flow and current OpenAI Codex models instead of a direct API key.

bash Codex OAuth setup

libre-claw auth codex-login
/model codex:gpt-5.6-sol --global

Anthropic

Direct Claude API access with Claude Opus 5 adaptive thinking, signed tool-loop continuity, and a one-million-token context window.

bash Anthropic setup

libre-claw auth set-key anthropic
/model anthropic:claude-opus-5 --global

OpenAI

Direct OpenAI API access for GPT models.

bash OpenAI setup

libre-claw auth set-key openai
/model openai:gpt-5.5 --global

Local Ollama

Run against a local Ollama-compatible endpoint.

bash Local Ollama setup

ollama serve
/model ollama:qwen3.6:27b --global

Fallback slots

Fallbacks let a run switch providers after rate limits, downtime, or transient provider errors.

bash Fallbacks

/fallback list
/fallback set 1 openrouter:deepseek/deepseek-v4-flash
/fallback set 2 ollama:minimax-m3:cloud
/fallback clear 2

Global defaults

Use --global when the default should apply across the TUI, daemon, Telegram, and schedules.

bash Global model

/model openrouter:z-ai/glm-5.2 --global
/status
/models