Feature hasn't been suggested before.
Describe the enhancement you want to request
Feature hasn't been suggested before.
I have verified this feature I'm about to request hasn't been suggested before.
Describe the enhancement you want to request
When OpenCode starts, or when a valid LiteLLM API Key is configured, OpenCode should automatically detect and synchronize newly available models exposed by the LiteLLM /v1/models endpoint.
Currently, users need to manually update their opencode.json configuration whenever new models are added to the LiteLLM gateway. This becomes operationally expensive in enterprise environments where models are frequently added, removed, versioned, or re-routed internally.
The proposed enhancement is for OpenCode to periodically fetch the model catalog from the configured LiteLLM endpoint and automatically update the local provider model registry.
The LiteLLM gateway should become the source of truth for available models.
Expected behavior
If the configured provider is LiteLLM-compatible:
{
"provider": {
"litellm": {
"options": {
"baseURL": "https://gateway.company.com/v1"
}
}
}
}OpenCode should:
- Detect the configured LiteLLM/OpenAI-compatible endpoint
- Call:
- Compare remote models against the locally cached/provider models
- Automatically add newly discovered models
- Optionally remove stale/unavailable models
- Refresh the model list without requiring manual edits to
opencode.json
Suggested sync behavior
Possible triggers:
- On application startup
- On provider/API key change
- Manual refresh action
- Periodic background sync (optional)
Suggested interval:
- Every 15–60 minutes
- Or configurable
Enterprise use case
This is especially useful in enterprise AI Gateway environments using LiteLLM as a centralized model router.
Example:
OpenCode
↓
LiteLLM Gateway
↓
OpenAI / OpenRouter / Huawei MaaS / Anthropic / DeepSeek
In this architecture:
- Models are dynamically managed server-side
- Aliases may change frequently
- New approved models are constantly added
- Internal aliases are preferred over provider-native names
Example aliases:
finnet-llm-v1
finnet-llm-v2
finnet-code-v2
finnet-reasoning-v3
Without auto-sync, every workstation requires manual configuration updates.
Benefits
- Reduces operational overhead
- Eliminates manual model management
- Keeps OpenCode aligned with enterprise AI gateways
- Improves governance consistency
- Supports dynamic model rollout strategies
- Better developer experience
- Enables centralized model lifecycle management
Optional improvements
Potential optional enhancements:
- Cache synced models locally
- Show "Last synced at"
- Sync only models accessible to the current API key
- Allow disabling auto-sync
- Allow filtering by provider or access group
- UI button: "Refresh Models"
Additional context
LiteLLM already exposes standardized model discovery through:
This feature would make OpenCode significantly easier to operate in enterprise environments using LiteLLM as the centralized AI Gateway.
Feature hasn't been suggested before.
Describe the enhancement you want to request
Feature hasn't been suggested before.
I have verified this feature I'm about to request hasn't been suggested before.
Describe the enhancement you want to request
When OpenCode starts, or when a valid LiteLLM API Key is configured, OpenCode should automatically detect and synchronize newly available models exposed by the LiteLLM
/v1/modelsendpoint.Currently, users need to manually update their
opencode.jsonconfiguration whenever new models are added to the LiteLLM gateway. This becomes operationally expensive in enterprise environments where models are frequently added, removed, versioned, or re-routed internally.The proposed enhancement is for OpenCode to periodically fetch the model catalog from the configured LiteLLM endpoint and automatically update the local provider model registry.
The LiteLLM gateway should become the source of truth for available models.
Expected behavior
If the configured provider is LiteLLM-compatible:
{ "provider": { "litellm": { "options": { "baseURL": "https://gateway.company.com/v1" } } } }OpenCode should:
opencode.jsonSuggested sync behavior
Possible triggers:
Suggested interval:
Enterprise use case
This is especially useful in enterprise AI Gateway environments using LiteLLM as a centralized model router.
Example:
In this architecture:
Example aliases:
Without auto-sync, every workstation requires manual configuration updates.
Benefits
Optional improvements
Potential optional enhancements:
Additional context
LiteLLM already exposes standardized model discovery through:
This feature would make OpenCode significantly easier to operate in enterprise environments using LiteLLM as the centralized AI Gateway.