Models

Press * to open the model and agent switcher. The agents tab shows everything you can connect to. The remaining tabs are model providers — switch the inference source without changing anything about your agent.

OpenRouter

The openrouter tab gives access to the full OpenRouter model catalogue — hundreds of models from every major provider, billed through a single OpenRouter account.

:keys set openrouter

Enter your OpenRouter API key when prompted. Models are grouped by provider in the palette.

TUI model switcher on the openrouter tab showing models grouped by provider

OpenAI / Codex

The openai tab lists OpenAI models including GPT-4 variants and o-series reasoning models. Codex subscribers see their available models here too.

:keys set openai

TUI model switcher on the openai tab showing GPT-4 and Codex models

Models are grouped by base model with reasoning levels nested underneath where applicable.

Ollama (local inference)

The local tab connects to Ollama running on your machine. Pull a model, select it, and inference runs entirely on your hardware — no network calls to external providers.

ollama pull qwen2.5-coder:32b

Then select it from the local tab. The TUI connects to Ollama on localhost:11434 automatically.

TUI model switcher on the local tab showing an Ollama model downloading

Managing Ollama from the TUI

The :ollama command group manages the Ollama server and installed models without leaving the TUI:

Command	What it does
`:ollama ps`	Show models currently loaded in VRAM
`:ollama unload <model>`	Unload a model from VRAM
`:ollama remove <model>`	Delete a model from disk
`:ollama update`	Re-pull all installed models to latest
`:ollama stop`	Stop the Ollama server

TUI command palette showing the ollama subcommand group with ps, unload, remove, update, and stop

Axon Cloud

The axon tab lists models backed by Axon Cloud inference — no key required beyond your Axon account. These run on managed infrastructure with the full Cognos cognitive loop.

Bring your own key

Keys are stored locally in ~/.axon/ and never sent to Axon servers. To see what's configured:

:keys show

To remove all saved keys:

:keys clear

Each provider's key is set independently. You can have OpenRouter and OpenAI configured simultaneously and switch between them freely.