Repository navigation
feat: add Cheaper Inference inference provider - #1623
Merged
mldangelo-oai merged 2 commits intoOct 19, 2026
Merged
Conversation
mldangelo-oai
approved these changes
Oct 19, 2026
mldangelo-oai
left a comment
Collaborator
There was a problem hiding this comment.
The new provider uses the existing explicit provider/model selection and credential-isolation paths. The preset matches the provider's documented Responses API, and synthetic checks cover API-key selection, MCP forwarding, and fresh/resumed Deep Scan workers without changing the default provider. I found no actionable correctness or simplification issues.
Merged
3 tasks done
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR adds Cheaper Inference as an inference provider for scans.
It uses the same pattern as the OpenRouter and Fireworks AI providers (#245).
Cheaper Inference is an OpenAI-compatible LLM gateway.
One API key gives access to models from several labs.
Usage:
Changes
CHEAPERINFERENCE_CODEX_PROVIDERtoEXTERNAL_CODEX_PROVIDERS(Responses wire API).--provideraccepts the new valuecheaperinference. The default staysopenai.CHEAPER_INFERENCE_API_KEYlike the other external provider keys (selection, isolation, plugin helper secrets, MCP env passthrough).unsetline.Details
Files:
sdk/typescript/src/config.ts,cli.ts,api.ts,runtime.tsplugins/codex-security/.mcp.json,plugins/codex-security/mcp-app/tests/test_mcp_app_smoke.tsREADME.md,sdk/typescript/docs/cli.md,docs/bedrock.mdapi-preflight-config,api,api-events,cli,runtime(new provider rows in the existing provider tables)Testing
tsc --noEmit(SDK and MCP app): pass.pnpm run build:ci: pass.prettier --checkon changed files andnode scripts/check-markdown.mjs: pass.node .github/scripts/check_plugin_source_compatibility.mjsand itsnode --testsuite: pass.bun teston the five changed SDK test files: the new rows pass where the OpenRouter and Fireworks rows pass. Tests that need_bundled_pluginfail the same way onmainwithout the native payload.gpt-5.4-minireturned a reply through the Codex SDK withEXTERNAL_CODEX_PROVIDERS.cheaperinferenceas the model provider.Risk and rollout
Additive change.
Existing providers and defaults do not change.
The new key is only read when
--provider cheaperinferenceis selected.Public disclosure review