[heydryft]

Pull Request

Description

Adds Runcrate as a preset model provider in the add-platform picker. Runcrate serves an OpenAI-compatible inference API at https://api.runcrate.ai/v1 (/chat/completions with streaming and tools, /models), so the entry uses the existing custom platform with that base URL and needs no new code paths.

Users pick "Runcrate" in the list, paste a key from the API keys page (keys look like rc_live_...) and fetch models. Model ids are vendor/model, for example deepseek-ai/DeepSeek-V4.1-Flash, zai-org/GLM-5.3-Flash or Qwen/Qwen3.8-Flash; the full list is on the model pricing page.

The change mirrors the shape of #4275: one entry appended to MODEL_PLATFORMS, a logo bundled under renderer/assets/logos/ (the backend logo service has no asset for it, so it is imported locally), and unit tests for the entry.

Files:

• packages/desktop/src/renderer/utils/model/modelPlatforms.ts: new Runcrate entry (platform: 'custom', base_url: 'https://api.runcrate.ai/v1').
• packages/desktop/src/renderer/assets/logos/runcrate.svg: 24x24 logo.
• tests/unit/renderer/modelPlatforms.test.ts: tests that the entry points at the OpenAI-compatible endpoint and is returned by getPresetProviders() with a logo.

Related Issues

• None

Type of Change

• fix: Bug fix (non-breaking change which fixes an issue)
• feat: New feature (non-breaking change which adds functionality)
• perf: Performance improvement
• refactor: Code restructuring (no behavior change)
• Breaking change (fix or feature that would break existing functionality)
• docs: Documentation update

Atomic PR Checklist (Rule 1)

• This PR contains exactly one feature or bug fix that cannot be further decomposed
• The PR title follows Conventional Commit format: <type>(<scope>): <subject> (English)

Local Checks (Rule 3)

• bun run format: formatting passes (bunx oxfmt --check on the touched files: all use the correct format)
• bun run lint: no lint errors (bunx oxlint on the touched files: 0 warnings, 0 errors)
• bunx tsc --noEmit: no type errors (exit 0)
• bunx vitest run: tests pass (bunx vitest run tests/unit/renderer/modelPlatforms.test.ts: 5 passed)
• i18n validated: N/A, no locale or i18n changes (the provider name is a brand and carries no i18nKey, like the other preset providers)
• New/changed user-facing text uses i18n keys (no hardcoded strings)

Runtime Verification

• Verified on macOS
• Verified on Windows
• Verified on Linux
• I have performed a self-review of my own code

The desktop app was not launched. Instead, the request path the app uses for a custom platform was run directly: ClientFactory.createRotatingClient with platform: 'custom' and base_url: 'https://api.runcrate.ai/v1', then createChatCompletion and models.list(), from a throwaway script with bun run on macOS:

• deepseek-ai/DeepSeek-V4.1-Flash: chat completion returned, finish_reason=stop, text "pong"
• zai-org/GLM-5.3-Flash: chat completion returned, finish_reason=stop, text "pong"
• models.list(): 197 models

Screenshots

None.

Additional Context

This PR was prepared with AI assistance: an AI coding agent wrote the change and ran the checks above. The commit carries no AI signature.

Disclosure: I work on Runcrate.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!