feat(provider): add AI/ML API provider

[D1m7asis]

Summary

This PR adds first-class AI/ML API support as an OpenAI-compatible provider across setup, runtime, model discovery, and provider profile flows. From #835

What changed

• added a dedicated AI/ML API provider preset in the provider picker
• added AI/ML API setup docs and README onboarding instructions
• introduced provider-specific helpers in src/providers/aimlapi/
• added support for AIMLAPI_API_KEY as a provider-specific credential
• synced AIMLAPI_API_KEY into OpenAI-compatible runtime flow when OPENAI_BASE_URL points to AI/ML API
• added AI/ML API attribution headers for requests sent to api.aimlapi.com
• improved /models discovery for AI/ML API by:
  • filtering to chat-completions models
  • deduplicating model IDs
  • mapping metadata into user-friendly labels/descriptions
• updated provider summary, validation, bootstrap, profile persistence, and secret handling to recognize AI/ML API
• added focused tests for provider preset rendering, credential fallback, attribution headers, model discovery, and preset defaults

Why it changed

OpenClaude already supports OpenAI-compatible providers, but AI/ML API needed a smoother first-class setup path and provider-aware behavior.

This change makes AI/ML API easier to configure, ensures the correct credential is used without forcing users to rename env vars, and improves model selection UX by showing only relevant chat-capable models with better metadata.

Impact

• user-facing impact:

  • users can now select AI/ML API directly in /provider
  • users can configure AI/ML API with AIMLAPI_API_KEY instead of only OPENAI_API_KEY
  • model discovery is cleaner and more useful for AI/ML API users
  • README and dedicated setup docs now include explicit AI/ML API onboarding

• developer/maintainer impact:

  • adds a provider-specific module for AI/ML API behavior
  • keeps OpenAI-compatible support generic while allowing targeted handling for AI/ML API
  • extends provider profile, validation, bootstrap, and secret flows with explicit AI/ML API support
  • adds regression coverage for the new provider path

Testing

• bun run build
• bun run smoke
• focused tests:
  • src/providers/aimlapi/index.test.ts
  • src/components/ProviderManager.test.tsx
  • src/services/api/openaiShim.test.ts
  • src/utils/providerDiscovery.test.ts
  • src/utils/providerProfiles.test.ts

Notes

• provider/model path tested:

  • /provider → AI/ML API
  • env-based setup with:
    • CLAUDE_CODE_USE_OPENAI=1
    • AIMLAPI_API_KEY=...
    • OPENAI_BASE_URL=https://api.aimlapi.com/v1
    • OPENAI_MODEL=gpt-4o
  • model discovery via GET /models against https://api.aimlapi.com/v1

• screenshots attached (if UI changed):

  • provider picker with AI/ML API preset

• optional: provider summary showing AI/ML API

• follow-up work or known limitations:
  • current model discovery intentionally focuses on chat-completions models used by OpenClaude's coding workflow
  • other AI/ML API modalities are documented but are outside the current provider loop

[Copilot]

Pull request overview

This PR adds first-class support for the AI/ML API service as an OpenAI-compatible provider, spanning provider selection UX, env/profile handling, runtime request behavior, and model discovery.

Changes:

• Adds an AI/ML API provider preset, plus env/profile support for AIMLAPI_API_KEY as a first-class credential.
• Adds AI/ML API–specific runtime behavior (attribution headers + credential fallback/sync) when targeting api.aimlapi.com.
• Improves /models discovery for AI/ML API (chat-completions filtering, dedupe, and metadata mapping) and adds docs/tests for the new provider.

Reviewed changes

Copilot reviewed 19 out of 19 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
src/utils/providerValidation.ts	Accepts AIMLAPI_API_KEY as an alternative credential when base URL is AI/ML API and improves missing-key messaging.
src/utils/providerSecrets.ts	Treats AIMLAPI_API_KEY as a secret for sanitization/redaction.
src/utils/providerProfiles.ts	Adds aimlapi preset defaults and ensures profiles clear/apply AIMLAPI_API_KEY where appropriate.
src/utils/providerProfiles.test.ts	Adds tests for AI/ML API preset defaults and credential fallback behavior.
src/utils/providerProfile.ts	Adds AIMLAPI_API_KEY to profile env/secret key sets and associated types.
src/utils/providerDiscovery.ts	Adds AI/ML API attribution headers and model-option discovery with metadata mapping; introduces listOpenAICompatibleModelOptions.
src/utils/providerDiscovery.test.ts	Tests AI/ML API model option mapping, filtering, deduplication, and headers.
src/services/api/providerConfig.ts	Allows additional model option caching for AI/ML API base URLs (in addition to local providers).
src/services/api/openaiShim.ts	Syncs AIMLAPI credentials into OpenAI flow for AI/ML API base URLs and adds attribution headers.
src/services/api/openaiShim.test.ts	Tests AI/ML API attribution headers and AIMLAPI_API_KEY fallback behavior.
src/services/api/bootstrap.ts	Uses model-option discovery (with metadata) and attempts AIMLAPI credential fallback for model discovery.
src/providers/aimlapi/index.ts	New provider module defining constants, base URL detection, attribution headers, credential helpers, preset defaults, and model mapping.
src/providers/aimlapi/index.test.ts	Unit tests for the new AI/ML API provider helper module.
src/components/ProviderManager.tsx	Adds AI/ML API preset to the provider picker.
src/components/ProviderManager.test.tsx	Verifies AI/ML API appears in the preset picker output.
src/commands/provider/provider.tsx	Updates provider summary labeling to identify AI/ML API specifically.
README.md	Adds onboarding instructions and links for AI/ML API setup.
docs/aimlapi-setup.md	Adds dedicated AI/ML API setup documentation.
.env.example	Documents AI/ML API env var configuration option.

Comments suppressed due to low confidence (1)

src/services/api/bootstrap.ts:153

• getAdditionalModelOptionsCacheScope() now allows AI/ML API (non-local) to use this discovery path, but the log message still says "Local OpenAI model discovery failed". This makes debugging confusing when the failing request was against a remote AI/ML API endpoint; consider updating the wording to reflect "OpenAI-compatible" (or include the resolved base URL/provider label).

  if (models === null) {
    logForDebugging('[Bootstrap] Local OpenAI model discovery failed')
    return null

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[kevincodex1]

hello thank you for this. some test failed kindly take a look

[D1m7asis]

hello thank you for this. some test failed kindly take a look

Thanks!

[Vasanthdev2004]

Thanks for the PR. I reviewed the current head as a full review.

Verdict: Needs changes

Blocking issue:

1. src/services/api/providerConfig.ts and src/services/api/bootstrap.ts now route AI/ML API through the bootstrap additional-model discovery path, which means startup/bootstrap can make an authenticated background GET https://api.aimlapi.com/v1/models request.
2. The existing nonessential-traffic/privacy guard is only applied in fetchBootstrapAPI(), not in the OpenAI-compatible bootstrap branch. Before this change that branch was effectively for local providers; on the current head it now reaches a remote third-party host even when nonessential traffic is disabled. That is a trust-boundary/privacy regression.

Non-blocking notes:

• I did not find a separate blocker in the provider preset wiring, host detection, header injection, or AIMLAPI/OpenAI key fallback logic.
• Saved startup profiles still serialize through the generic OPENAI_API_KEY path rather than preserving AIMLAPI_API_KEY end-to-end.
• I am also not seeing any GitHub check-runs or statuses on the current head, so there is no CI signal attached to this review.

Happy to re-review once the AI/ML API model discovery path is either explicitly user-driven or gated behind the existing nonessential-traffic/privacy rules.

[kevincodex1]

hello @D1m7asis thank you for this, we pause the provider integration to give way to this PR: #910 once its merged you'll need to rewrite the approach. this will be more optimized approach

[D1m7asis]

okay thanks. I`ll wait

[jatmn]

#910 merged, please rebase and fix conflicts.
note: providers changed alot

[Vasanthdev2004]

I would not re-review this current head yet because #910 has now landed and changed the provider architecture quite a lot.

The useful next step is to rebase/rework this against current main using the descriptor-backed provider route shape instead of the older provider-specific wiring. After that, the main thing I would re-check is the earlier trust-boundary concern: AI/ML API model discovery should not introduce authenticated remote/background traffic unless it is user-driven or gated by the existing nonessential-traffic/privacy rules.

Once the branch is rebased and conflicts are resolved, happy to do a fresh review on the new shape.