Greptile Summary

This PR adds an optional Model field to the AI provider configuration, letting users pin a specific model ID per provider. When set, the model is injected as the default in getAllModels() and used as a fallback in streamChat when the client does not specify one.

• Server (index.ts): LlmProviderSetup gains model?: string; getAllModels marks the configured model as isDefault (or prepends it when absent from the provider's static list); a new getProviderSetupByType export exposes the setup config for a provider type.
• Server (llm_chat.ts): streamChat resolves effectiveModel = config.model || setup?.model so a provider-level default is applied when the client omits a model.
• Client: AddProviderModal gains a trimmed, optional model input; ProviderList shows the configured model (or a "Default" placeholder) in the table; translations are added for EN and RU.

Confidence Score: 5/5

Safe to merge; the feature is additive and purely opt-in — omitting the new field leaves all existing behavior unchanged.

All changed paths are guarded by optional fields with clear fallbacks. The model-injection logic in getAllModels correctly handles both the 'model already in list' and 'model not in list' cases and strips isDefault from other entries. The streamChat fallback chain (config.model → setup.model → provider default → error) is well-structured. The one gap (handler-level test for the setup.model fallback) is a coverage nit that does not affect correctness.

No files require special attention; the test gap in llm_chat.spec.ts is the only item worth a follow-up.

Important Files Changed

sequenceDiagram
    participant Client
    participant streamChat
    participant getProviderByType
    participant getProviderSetupByType
    participant Provider

    Client->>streamChat: "POST /llm/chat {config: {provider, model?}}"
    streamChat->>getProviderByType: getProviderByType(config.provider)
    getProviderByType-->>streamChat: LlmProvider instance
    streamChat->>getProviderSetupByType: getProviderSetupByType(config.provider)
    getProviderSetupByType-->>streamChat: "LlmProviderSetup {model?}"
    Note over streamChat: effectiveModel = config.model || setup?.model
    streamChat->>Provider: "chat(messages, {...config, model: effectiveModel})"
    Provider-->>streamChat: StreamResult
    streamChat-->>Client: SSE chunks

_{Reviews (2): Last reviewed commit: "Update apps/client/src/translations/ru/t..." | Re-trigger Greptile}