LLM endpoint switching on the fly in telegram

[superhero75]

What problem does this solve?
Most llm endpoints have rate limit, it is difficult to switch to another model/endpoint on the road.
Proposed solution
in telegram, /models and /model to switch endpoints or models (openclaw feature)
Alternatives considered
n/a
Scope estimate
No idea

Would you like to implement this?
No

[qhkm]

Hey @superhero75 — good news, this is already implemented! 🎉

You can use the /model command in Telegram right now:

• /model — show current model
• /model list — show available models grouped by provider (with [ok]/[warn] API key status)
• /model openai:gpt-5.1 — switch to a specific provider + model
• /model gpt-5.1 — switch model (keep current provider)
• /model reset — clear override, back to default

It's per-chat and persisted across restarts via long-term memory. The provider registry is populated at startup with all configured providers, so /model switches both the model and the underlying LLM endpoint.

Let us know if you run into any issues with it!

[superhero75]

/model list isnt working properly as it isnt showing all the models which have been configured. so it is quite hard to type in the entire name of the model without the list of models

[qhkm]

Ah i see so you meant the autosuggest list. Sure will check that out