[herjarsa]

Problem

The GET /api/providers/[id]/models route fell back to the static PROVIDER_MODELS list when no per-connection cached discovery models existed. This meant:

• Obsolete models persisted in the UI even after sync removed them
• New models were unavailable until static PROVIDER_MODELS was manually updated
• Cross-connection gap: sync performed on Connection A wouldn't benefit Connection B's model listing

The catalog endpoint (GET /api/v1/models/catalog) already handled this by checking getAllSyncedAvailableModels() and skipping static providers. But the individual provider models route did not.

Solution

In GET /api/providers/[id]/models, before building the local catalog fallback:

1. Call `getSyncedAvailableModels(provider)$ — returns synced models across ALL connections for this provider
2. If any synced models exist, use them INSTEAD of PROVIDER_MODELS + getStaticModelsForProvider
3. If no synced models exist, fall through to static models as before

This ensures that once a sync completes for ANY connection of a provider, ALL subsequent model listing requests for any connection of that provider return the authoritative, up-to-date list.

Impact

• No new dependencies — uses existing getSyncedAvailableModels from src/lib/db/models
• Backward compatible — no synced models = same behavior as before
• Self-healing — stale/obsolete models disappear after first sync
• All providers — works for API-key, OAuth, managed, and self-hosted providers alike

Changes

• src/app/api/providers/[id]/models/route.ts: Import getSyncedAvailableModels, check before local catalog fallback

[gemini-code-assist]

Code Review

This pull request updates the GET handler in src/app/api/providers/[id]/models/route.ts to fetch and use synced available models from the database as the authoritative source instead of static models. The review feedback correctly points out that mapping the synced models to only { id, name } discards critical metadata like apiFormat and supportedEndpoints which are expected downstream by toLocalCatalogModels(), and provides a code suggestion to preserve these fields.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

By mapping allSynced to only { id, name }, we lose important metadata such as apiFormat and supportedEndpoints which are defined on SyncedAvailableModel and expected by toLocalCatalogModels(). We should preserve these fields by mapping them to LocalCatalogModel objects.

    let providerSyncedModels: LocalCatalogModel[] | null = null;
    try {
      const allSynced = await getSyncedAvailableModels(provider);
      if (Array.isArray(allSynced) && allSynced.length > 0) {
        providerSyncedModels = allSynced.map((m) => ({
          id: m.id,
          name: m.name || m.id,
          apiFormat: m.apiFormat,
          supportedEndpoints: m.supportedEndpoints,
        }));
      }
    } catch {
      // DB unavailable — fall through to static catalog
    }

[diegosouzapw]

Thanks @herjarsa! Using the synced model list as the authoritative local catalog (across all connections of a provider, even ones that didn't run the sync) is the right call — it keeps the model picker in sync with what actually exists upstream. I pushed a discriminating regression test (provider-models-route → "synced-models authoritative local catalog") that fails on the pre-fix route and passes with yours, so the behavior is now guarded. Merging into release/v3.8.10. 🙏