Hermes Agent does not see the custom model in Telegram, which does not necessarily mean that the model has not been successfully configured. Recent community feedback mentioned that the custom/Ollama model selector display issue is tracked separately; In other words, CLI can be listed with Telegram picker is not the same thing.
First, confirm whether the model can be used
- In the CLI, use the
hermes modelor configuration file to switch to the custom provider. - Send a normal task directly to confirm that the model can reply and call the underlying tool.
- If the CLI is available but the Telegram dropdown is not, it will be handled as a gateway display issue.
- The ad-hoc scenario is to fix the provider/model used by the gateway instead of relying on the picker.
Don't repeatedly switch to Open WebUI or local model services just because the Telegram list doesn't show up. Determine whether it is "unavailable" or "not displayed" first.
In a word: Telegram does not see the custom model, verify it with the CLI first, and then process the gateway model selector.
Official open source address: https://github.com/NousResearch/hermes-agent; Official document entrance: https://hermes-agent.nousresearch.com/.
