Skip to content

feat: add requesty and cheaper inference providers - #589

Merged
di-sukharev merged 8 commits into
di-sukharev:masterfrom
aiapienthusiast:cheaperinference-provider
Nov 13, 2026
Merged

di-sukharev merged 8 commits into
di-sukharev:masterfrom
aiapienthusiast:cheaperinference-provider

Conversation

@aiapienthusiast

@aiapienthusiast aiapienthusiast commented Sep 29, 2026 •

Copy link
Copy Markdown
Contributor

adds requesty and cheaper inference as optional openai-compatible providers, with model discovery and configurable inference endpoints. fresh global setups skip the legacy provider-key migration, and setup examples include an explicit model.

model discovery excludes non-chat models and cheaper inference models that only support the responses api.

includes the original requesty commits from #588.

validation:

  • npm run lint
  • npm run format:check
  • npm run test:unit -- --runInBand
  • npm run test:e2e

regression coverage includes the first commit with no migration records and chat endpoint compatibility.

binyangzhu000-sudo and others added 3 commits September 18, 2026 12:31
Signed-off-by: binyangzhu000-sudo <224954946+binyangzhu000-sudo@users.noreply.github.com>
Requesty is an OpenAI-compatible gateway, so RequestyEngine reuses
OpenAiEngine with the Requesty base URL (https://router.requesty.ai/v1)
by default while still allowing OCO_API_URL to override it (for example
the EU endpoint).

Registered alongside OpenRouter and OrcaRouter in the provider enum,
config validators, model list, key and billing URLs, recommended model,
engine factory, setup picker and model cache. Model discovery merges the
managed policies list with the full catalog, fetching each on its own.
Includes unit tests, README docs and the rebuilt bundle.
@di-sukharev

Copy link
Copy Markdown
Owner

fresh setup fails in migration 00. please fix this path and add a cli test.

binyangzhu000-sudo and others added 5 commits November 11, 2026 23:15
Setup and 'oco config set OCO_AI_PROVIDER=atlascloud' were rejected because the
validator kept its own hard-coded provider list. Derive it from
OCO_AI_PROVIDER_ENUM so new providers (and mlx, which was also missing) cannot
drift again, update the error messages, and add regression tests.

Signed-off-by: binyangzhu000-sudo <224954946+binyangzhu000-sudo@users.noreply.github.com>
@di-sukharev di-sukharev changed the title feat: add Cheaper Inference as a named AI provider feat: add requesty and cheaper inference providers Nov 13, 2026
@di-sukharev
di-sukharev merged commit e601776 into di-sukharev:master Nov 13, 2026
5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants