Add Telnyx model provider

[Oliver-Zimmerman]

Summary

Adds a Telnyx model provider plugin for Dify Official Plugins under models/telnyx.

This provider adds predefined Telnyx support for:

• LLM/chat completions via POST /v2/ai/chat/completions
• text embeddings via POST /v2/ai/openai/embeddings
• speech-to-text via POST /v2/ai/audio/transcriptions
• text-to-speech via Telnyx TTS

The predefined model lists were checked against Telnyx's authenticated read-only model/voice endpoints before opening this draft:

• LLM repo/live match: 18 / 18
• Embedding repo/live match: 3 / 3
• Selected TTS voices all exist in the live voice catalog: 12 / 3329

Local Dify install/runtime smoke has been completed; screenshots are posted in #3040 (comment).

Change Type

• Documentation / non-plugin change
• Non-LLM plugin (tools, extensions, datasource, etc.)
• LLM plugin

Screenshots / Videos

Before	After
Telnyx was not available as an official model provider.	Telnyx plugin installs locally, credentials validate, and a Telnyx model returns a successful response through Dify.

LLM Plugin Checklist

Areas affected by this change (check all that apply)

• Message flow (system messages, user ↔ assistant turn-taking)
• Tool interaction flow (multi-round usage, Agent App and Agent Node)
• Multimodal input (images, PDFs, audio, video, etc.)
• Multimodal output (images, audio, video, etc.)
• Structured output (JSON, XML, etc.)
• Token consumption metrics
• Other LLM functionality (reasoning, grounding, prompt caching, etc.)
• New models / model parameter fixes

Version

• Bumped top-level version in manifest.yaml (not the one under meta)
• dify_plugin>=0.3.0,<0.6.0 is declared in pyproject.toml and locked in uv.lock (or kept in requirements.txt for legacy plugins without uv.lock) — SDK docs

Testing

• Local deployment — Dify version: 1.14.1
• SaaS (cloud.dify.ai)

Validation completed so far:

cd models/telnyx
uv run --python 3.12 python -m pytest tests -q
uv run --python 3.12 python -m compileall -q models provider tests main.py

Result:

8 passed, 2 warnings
YAML parse and AIModelEntity validation OK (24 model YAMLs)

Package smoke with the Dify plugin daemon CLI release asset:

cd dify-official-plugins
/tmp/dify-plugin-cli/dify-plugin plugin package models/telnyx --output_path /tmp/telnyx.difypkg
shasum -a 256 /tmp/telnyx.difypkg

Result:

plugin packaged successfully
/tmp/telnyx.difypkg
sha256: 97e9c58b28239ee9bd73af2b63a7e8240bff84c5f2478715b85a5fa1ae6da7f0

Live read-only inventory checks completed with a Telnyx API key, without logging secrets:

LLM repo/live match: True 18 18
Embedding repo/live match: True 3 3
TTS selected voices all live: True 12 of 3329

Runtime smoke completed:

• installed the generated .difypkg into a local Dify environment;
• configured Telnyx credentials in Dify;
• ran a minimal LLM smoke through Dify's model provider layer;
• added screenshots for plugin installation, provider configuration, and a successful chat response in Add Telnyx model provider #3040 (comment).

Optional embedding/TTS/STT UI smokes are still available as follow-up coverage if maintainers want the full provider surface exercised visually.

[gemini-code-assist]

Code Review

This pull request introduces the Telnyx model provider plugin, which adds support for LLM chat completions, text embeddings, text-to-speech, and speech-to-text services. The implementation includes shared HTTP utility classes, specific model providers, a wide range of predefined model configurations (including GPT-4o, Claude, and Llama models), and a comprehensive test suite. Review feedback highlighted critical issues in the LLM implementation, specifically regarding incorrect tool mapping logic in the request payload and a streaming response handler that failed to capture tool calls when text content was missing from a chunk.

[Oliver-Zimmerman]

Screenshots / Runtime smoke

Tested against a local Dify deployment using the packaged PR plugin.

Validation performed:

• Built /tmp/telnyx.difypkg from models/telnyx
• Installed the local package into Dify
• Configured Telnyx provider credentials with the API key masked by Dify
• Ran a chat completion through Dify using the Telnyx provider

The successful LLM smoke below uses moonshotai/Kimi-K2.6 through the Telnyx provider.

1. Telnyx plugin installed

2. Telnyx provider configured

3. Telnyx LLM chat success

[Oliver-Zimmerman]

@RockChinQ Would be great to have this reviewd / merged.

[Oliver-Zimmerman]

@crazywoola would love to get this integrated.

[Oliver-Zimmerman]

@orangeclk