Add Telnyx as an inference provider

[jamestwhedbee]

Registers Telnyx as a new Inference Provider for the JS client as step 2 (JS Client Integration) of the provider registration guide.

Telnyx exposes an OpenAI-compatible chat completions API. This PR adds conversational-task support only for now, using the direct provider base URL https://api.telnyx.com/v2/ai/openai.

One provider-specific detail: Telnyx's OpenAI-compatible base URL already includes the version segment, so the conversational helper overrides the default route and uses chat/completions instead of v1/chat/completions.

Validation:

• pnpm --filter @huggingface/inference format
• pnpm --filter @huggingface/inference lint
• pnpm --filter @huggingface/inference build
• pnpm --filter @huggingface/inference test
• pnpm --filter @huggingface/inference check

---

Note

Low Risk
Additive provider integration following existing conversational patterns; no changes to auth, billing, or other providers’ behavior.

Overview
Adds Telnyx as a new inference provider in @huggingface/inference, with conversational / chat completion support only.

A new TelnyxConversationalTask targets https://api.telnyx.com/v2/ai/openai and overrides the default OpenAI-style path to use chat/completions (not v1/chat/completions) because Telnyx’s base URL already includes the API version. telnyx is registered in provider types, hub org mapping, hardcoded model mapping, and getProviderHelper. The README provider and supported-models lists are updated, and a unit test asserts the resolved base URL and route for direct provider-key calls.

^{Reviewed by Cursor Bugbot for commit 6c590f6. Bugbot is set up for automated code reviews on this repo. Configure here.}