Problem or Use Case
When using Hermes from Telegram or other gateway surfaces, there is no clear built-in way to list the models available for the current provider. /model is useful for switching models, but it does not reliably act as a discoverable “list models” command in messaging contexts.
This makes it harder to:
- discover valid model names
- switch providers confidently
- use external or custom providers without trial-and-error
- work efficiently from Telegram, where there is no rich interactive model picker
Right now, the workaround is to manually query provider endpoints like hermes model outside the channel, which breaks the flow and is not obvious to users. A dedicated /models command would solve this by making model discovery a first-class part of the interface.
Proposed Solution
Add a dedicated /models slash command that lists available models for the current provider by default, and optionally for a specific provider when one is passed as an argument.
Proposed behavior:
/models → list models for the current active provider/models <provider> → list models for that provider- examples:
/models/models openrouter/models anthropic/models openai
Output behavior:
- plain text, compact, messaging-friendly
- one model per line
- mark the currently active model if possible
- if the provider supports live model discovery, fetch from its
/models endpoint
- if live discovery is not supported, fall back to Hermes’ built-in catalog for that provider
- if neither is available, return a clear error message
For custom providers:
- support named custom providers too, for example:
/models custom:lmstudio/models custom:work
- for OpenAI-compatible custom endpoints, fetch from
<base_url>/models
Telegram/Discord UX:
- keep the response short and readable
- if the list is long, paginate or truncate with a “show more” pattern
- the command should work even if the Telegram slash-command menu is not refreshed, as long as the user types it manually
Nice to have:
- allow easy follow-up with
/model <provider:model>
- optionally group results by provider family or show small metadata like “current” or “default”
Alternatives Considered
No response
Feature Type
New tool
Scope
Small (single file, < 50 lines)
Contribution
Problem or Use Case
When using Hermes from Telegram or other gateway surfaces, there is no clear built-in way to list the models available for the current provider.
/modelis useful for switching models, but it does not reliably act as a discoverable “list models” command in messaging contexts.This makes it harder to:
Right now, the workaround is to manually query provider endpoints like
hermes modeloutside the channel, which breaks the flow and is not obvious to users. A dedicated/modelscommand would solve this by making model discovery a first-class part of the interface.Proposed Solution
Add a dedicated
/modelsslash command that lists available models for the current provider by default, and optionally for a specific provider when one is passed as an argument.Proposed behavior:
/models→ list models for the current active provider/models <provider>→ list models for that provider/models/models openrouter/models anthropic/models openaiOutput behavior:
/modelsendpointFor custom providers:
/models custom:lmstudio/models custom:work<base_url>/modelsTelegram/Discord UX:
Nice to have:
/model <provider:model>Alternatives Considered
No response
Feature Type
New tool
Scope
Small (single file, < 50 lines)
Contribution