AiHummer
English▾
Log in
v1.0.x▾
v1.0.xstable
{ }Swagger
Web UI/Models

Models

v1.0.x · updated 2026-07-08

The “Models” screen defines which language models agents use. Its core is a priority-ordered provider chain: AiHummer walks the list top to bottom until it gets a working answer, so you can put your primary provider first and fallbacks below.

The provider chain

Each provider is a card with a priority number and ↑/↓ arrows to reorder, its kind, model, base URL, a “key set” badge, an enabled/disabled status, and Test, “Edit” and “Delete” buttons.

Adding and configuring a provider

The add button opens a modal (the same one the setup wizard uses):

  • Kind (immutable after create): openai-compatible (any OpenAI-compatible endpoint), gateway (internal) or codex.
  • Name, base_url, api_key (password).
  • Model — first a text input with typical-model hints; after the key verifies, the model list is discovered from the provider’s /models (with a “custom” fallback).
  • An enabled checkbox and a Test button to check connectivity.

Codex (ChatGPT subscription)

The codex kind hides base_url/key and shows a device-auth connect panel: Connect / Disconnect / Refresh. This is how you use models through a ChatGPT subscription without a separate API key.

BYOK and LLM settings

  • BYOK (bring your own key) — a per-workspace override: provider, model, api_key, base_url; “Save” and “Clear” buttons.
  • LLM settings — the part of the settings catalog owned by this screen (temperature, timeouts, etc.).

How to use it

  1. Add your primary provider, click Test, enable it.
  2. Add a fallback below it in priority if needed.
  3. Restrict which models agents can see on Instructions → allowed models.

Next