feat: add Requesty as an OpenAI-compatible provider

[Thibaultjaigu]

Add Requesty as an OpenAI-compatible provider

This adds Requesty (https://router.requesty.ai/v1) as a first-class provider preset, mirroring how the existing gateway providers are wired.

How it fits the existing architecture

Providers in this repo are config-driven: VENDOR_PRESETS in src/lib/provider-catalog.ts is the single source of truth, and the quick-add UI, search-and-add, and model discovery are all generated from it. This PR adds a requesty preset there rather than adding any bespoke code path.

• src/lib/provider-catalog.ts — new requesty preset: base URL https://router.requesty.ai/v1, openai-compatible protocol, api_key auth, a few curated provider/model defaults, and meta doc links. Because the base URL ends in /v1 and the preset isn't catalog-only, Requesty automatically flows through the generic /v1/models search-and-add path (canReliablyFetchModels / discoverModels) — no dedicated catalog fetcher needed.
• src/lib/provider-icon-rule.ts — added "requesty" to ProviderIconKey plus a brand-scoped matcher (requesty.ai host / requesty name).
• src/components/settings/provider-presets.tsx — added a RequestyIcon and wired it into the icon maps (the model-selector already referenced a requesty logo slug; this makes the backend match it).

OpenRouter's preset uses the openrouter protocol (an Anthropic-skin for the Claude Code SDK). Requesty's gateway is OpenAI-compatible, so the faithful mirror is a branded openai-compatible preset, exactly like the existing branded OpenAI-compatible presets in the catalog.

Verification

• tsc --noEmit: clean before and after.
• test:unit: 4540 pass / 0 fail.
• ESLint on touched files: 0 errors.
• Live test against the real endpoint: chat completion with openai/gpt-4o-mini → 200; GET /v1/models → 200 with 596 models (context_window / supports_tool_calling fields present).

I work at Requesty. This mirrors the existing OpenRouter provider as closely as possible. Happy to adjust or close it if it's not a fit.

[vercel]

@Thibaultjaigu is attempting to deploy a commit to the op7418's projects Team on Vercel.

A member of the Team first needs to authorize it.