You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Using Opper with Zero today means setting up custom-openai-compatible and entering the endpoint, key variable and model by hand. I would like to propose a named opper preset, built the same way as the Atlas Cloud preset in #784.
Opper is an EU-hosted AI gateway: one OpenAI-compatible API and one key for 700+ models from 30+ providers, including Claude, GPT, Gemini, DeepSeek and Kimi. It is hosted in the EU (AWS Stockholm), and requests can be pinned to EU-hosted regions.
Proposed scope
One descriptor using the existing openAICompat helper:
Bearer auth on the existing OpenAI-compatible chat, tool and streaming runtime. No new transport, dependency, OAuth flow or default-provider change.
Placed in the aggregator cluster directly after openrouter, as asked in the review on feat(providers): add AI/ML API preset #402. Not badged as recommended and not above the first-party providers.
No attribution, referral or partner headers. Requests go through plain CopyHeaders like every other preset.
Curated picker entries in internal/providermodelcatalog/catalog.go: claude-sonnet-4-6 (default), gpt-5.5, gemini-3.8-flash, deepseek-v4-pro and kimi-k3. All five are pooled ids in Opper's public catalogue (GET https://api.opper.ai/v3/models), checked on 2026-10-02.
Remote model list: ModelsDevProviderID falls through to the catalog id, and models.dev already lists provider opper, so remote.go needs no mapping line.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Using Opper with Zero today means setting up
custom-openai-compatibleand entering the endpoint, key variable and model by hand. I would like to propose a namedopperpreset, built the same way as the Atlas Cloud preset in #784.Opper is an EU-hosted AI gateway: one OpenAI-compatible API and one key for 700+ models from 30+ providers, including Claude, GPT, Gemini, DeepSeek and Kimi. It is hosted in the EU (AWS Stockholm), and requests can be pinned to EU-hosted regions.
Proposed scope
One descriptor using the existing
openAICompathelper:openrouter, as asked in the review on feat(providers): add AI/ML API preset #402. Not badged as recommended and not above the first-party providers.CopyHeaderslike every other preset.internal/providermodelcatalog/catalog.go:claude-sonnet-4-6(default),gpt-5.5,gemini-3.8-flash,deepseek-v4-proandkimi-k3. All five are pooled ids in Opper's public catalogue (GET https://api.opper.ai/v3/models), checked on 2026-10-02.ModelsDevProviderIDfalls through to the catalog id, and models.dev already lists provideropper, soremote.goneeds no mapping line.Files, mirroring #784:
internal/providercatalog/catalog.gointernal/providercatalog/catalog_test.go(expectedCatalogIDs, theTransportOpenAICompatorder list, aTestOpperDescriptor)internal/providermodelcatalog/catalog.gointernal/providermodelcatalog/catalog_test.go(a case inTestModelsAreProviderScoped)No README or promotional changes.
If this is welcome, I will file the feature issue and open the PR once it has
issue-approved.Disclosure: I work at Opper.
All reactions