feat: add Telnyx as OpenAI-compatible provider

[gbattistel]

Summary

Adds Telnyx as an OpenAI-compatible provider in LiteLLM via the providers.json configuration system.

Telnyx provides an OpenAI-compatible Inference API for LLM access to hosted models (Kimi-K2.6, GLM-5.1, MiniMax-M2.7, and more).

Changes

• litellm/llms/openai_like/providers.json — Add Telnyx provider entry with base_url: https://api.telnyx.com/v2/ai and api_key_env: TELNYX_API_KEY
• litellm/types/utils.py — Add TELNYX = "telnyx" to the LlmProviders enum
• provider_endpoints_support.json — Add Telnyx endpoint support matrix (chat completions + embeddings)
• tests/test_litellm/llms/openai_like/test_telnyx_provider.py — Unit tests for provider config, resolution, and enum
• docs/my-website/docs/providers/telnyx.md — Provider documentation
• docs/my-website/sidebars.js — Add Telnyx to providers sidebar

Usage

import os
import litellm

os.environ["TELNYX_API_KEY"] = "your-api-key"

response = litellm.completion(
    model="telnyx/moonshotai/Kimi-K2.6",
    messages=[{"role": "user", "content": "Hello!"}],
)

Testing

All 7 unit tests pass:

• ✅ Provider exists in JSONProviderRegistry
• ✅ Provider config (base_url, api_key_env)
• ✅ Dynamic config generation
• ✅ Provider resolution (telnyx/model-name → provider=telnyx)
• ✅ Supported params
• ✅ ProviderConfigManager integration
• ✅ LlmProviders enum entry

References

• Telnyx Inference API Docs: https://developers.telnyx.com/docs/inference/getting-started
• Telnyx Available Models: https://developers.telnyx.com/docs/inference/models
• LiteLLM Provider Registration: https://docs.litellm.ai/docs/provider_registration/

[veria-ai]

Low: No security issues found

This PR adds Telnyx as a new OpenAI-compatible provider. The changes are limited to documentation, a JSON configuration entry (base URL and env var name for the API key), an enum addition, a sidebar entry, endpoint support metadata, and unit tests. No new code paths handle user input, authentication, or data processing — the provider is wired through the existing JSON-based provider system.

---

Status: 0 open
Risk: 1/10

Posted by Veria AI · 2026-04-24T00:21:32.220Z

[greptile-apps]

Greptile Summary

This PR registers Telnyx as an OpenAI-compatible provider by adding a providers.json entry, the corresponding LlmProviders.TELNYX enum value, endpoint support metadata, unit tests, and documentation. The implementation follows the established JSON-provider pattern used by other providers in this directory (e.g. veniceai, scaleway). Previously flagged issues (sidebar duplicate, fragile pytest import guard) have both been resolved.

Confidence Score: 5/5

Safe to merge — all changes are additive and follow established patterns with no backwards-incompatible modifications.

All findings are P2 style suggestions. The core implementation (providers.json, enum, endpoint support matrix) is correct and consistent with existing JSON-registered providers. Previously reported issues are addressed. No network calls in tests, no security concerns, no breaking changes.

No files require special attention.

Important Files Changed

Filename	Overview
litellm/llms/openai_like/providers.json	Adds telnyx entry with correct base_url and api_key_env; consistent with existing JSON-provider pattern
litellm/types/utils.py	Adds TELNYX = "telnyx" to LlmProviders enum; clean placement and consistent with other entries
provider_endpoints_support.json	Adds Telnyx endpoint support matrix with chat_completions and embeddings enabled; all other endpoints correctly set to false
tests/test_litellm/llms/openai_like/test_telnyx_provider.py	Unit tests cover provider registration, config values, resolution, and enum; no real network calls; uses sys.path.insert workspace hack and an if __name__ == "__main__" runner block
docs/my-website/docs/providers/telnyx.md	Provider documentation with accurate base URL, model table, and usage examples; the OpenAI SDK example correctly uses the /openai-suffixed endpoint per Telnyx's quickstart docs
docs/my-website/sidebars.js	Adds a single providers/telnyx entry in alphabetical order; previously reported triple-duplicate is resolved

Sequence Diagram

sequenceDiagram
    participant User
    participant LiteLLM
    participant JSONProviderRegistry
    participant TelnyxAPI

    User->>LiteLLM: completion(model="telnyx/moonshotai/Kimi-K2.6")
    LiteLLM->>JSONProviderRegistry: exists("telnyx")?
    JSONProviderRegistry-->>LiteLLM: true (providers.json entry)
    LiteLLM->>JSONProviderRegistry: get("telnyx")
    JSONProviderRegistry-->>LiteLLM: {base_url, api_key_env}
    LiteLLM->>LiteLLM: resolve api_base = "https://api.telnyx.com/v2/ai"
    LiteLLM->>LiteLLM: resolve api_key from TELNYX_API_KEY env
    LiteLLM->>TelnyxAPI: POST /v2/ai/chat/completions
    TelnyxAPI-->>LiteLLM: OpenAI-compatible response
    LiteLLM-->>User: ModelResponse

Loading

_{Reviews (3): Last reviewed commit: "fix: remove fragile pytest import guard ..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!