Telnyx AI

This repo is the one-stop shop for AI Agents and AI-first developers building with Telnyx.

[!NOTE] This repository is a work in progress under active development. We are continuously improving based on testing and feedback. Contributions and feedback encouraged!

Table of contents

• Telnyx Plugins - Install the Telnyx plugin for Claude Code, Cursor, or Gemini CLI to give your coding assistant Telnyx MCP server access and Telnyx Agent Skills.

• Agent Toolkit - integrate Telnyx APIs with popular agent frameworks including OpenAI's Agent SDK, LangChain, CrewAI, and Vercel's AI SDK through function calling — available in Python and TypeScript.

• Agent Skills - give AI coding assistants accurate, up-to-date context about Telnyx APIs and SDKs.

• Agent CLI - provision and build on Telnyx infrastructure in a single command.

• Model Context Protocol (MCP) - use Telnyx's generic API MCP proxy or app-layer MCP Apps.

• Guides - step-by-step tutorials for common workflows

Plugins and Extenstion

Install the unified Telnyx plugin to give your AI coding assistant Telnyx MCP server access and 228 Agent Skills covering messaging, voice, numbers, AI, IoT, WebRTC, Twilio migration, and more.

Empowers coding agents to generate correct, production-ready code without relying on pre-training or fragile doc retrieval.

Claude Code Plugin

Step 1. Add the Telnyx marketplace (one-time setup):

/plugin marketplace add team-telnyx/ai

Step 2. Install the plugin:

/plugin install telnyx@telnyx

Gemini CLI extension

 gemini extensions install https://github.com/team-telnyx/ai

OpenCode

Install the Telnyx plugin for OpenCode to add Telnyx as a model provider with automatic auth and a TUI for managing hosted models.

# Local (current project only)
opencode plugin @telnyx/opencode

# Global (all projects)
opencode plugin -g @telnyx/opencode

See plugins/opencode/README.md for full setup and configuration.

Cursor

[!NOTE] Note: Our Cursor Marketplace listing is pending.

In the meantime, install skills via the Skills CLI.

Add the Telnyx MCP server to your project's .cursor/mcp.json:

  {                                                         
    "mcpServers": {
      "telnyx": {
        "type": "http",
        "url": "https://api.telnyx.com/v2/mcp"
      }
    }
  }

Harnesses

Finalized Telnyx harness plugin repositories for OpenClaw and Hermes integrations:

OpenClaw

Hermes

Hermes voice-call support is in progress and will be added once finalized.

Agent Toolkit

Integrate Telnyx APIs with popular agent frameworks through function calling — available in Python and TypeScript.

Python

pip install telnyx-agent-toolkit

from telnyx_agent_toolkit.openai.toolkit import TelnyxAgentToolkit

toolkit = TelnyxAgentToolkit(
    api_key="KEY_...",
    configuration={
        "actions": {
            "messaging": {"send_sms": True},
            "numbers": {"search_phone_numbers": True, "buy_phone_number": True}
        }
    }
)

tools = toolkit.get_openai_tools()

Works with OpenAI's Agent SDK, LangChain, and CrewAI. See Python docs for full usage and examples.

TypeScript

npm install @telnyx/agent-toolkit

import { TelnyxAgentToolkit } from "@telnyx/agent-toolkit/langchain";

const toolkit = new TelnyxAgentToolkit(process.env.TELNYX_API_KEY!, {
  configuration: {
    actions: {
      messaging: { send_sms: true },
      numbers: { search_phone_numbers: true, buy_phone_number: true },
    },
  },
});

const tools = toolkit.getLangChainTools();

Works with LangChain and Vercel's AI SDK. See TypeScript docs for full usage. for the full list of commands and options.

Agent Skills

Install individual skills for your coding assistant via the Skills CLI:

npx skills add team-telnyx/ai --skill <SKILL> --agent <AGENT>

[!NOTE] See Skills for full install instrcuctions and comprehensive list of available skills

Agent CLI

Composite commands that reduce multi-step Telnyx workflows to a single command. Built for AI agents and developers who want to provision infrastructure without orchestrating multiple API calls.

telnyx-agent setup-sms        # Buy number + create messaging profile + assign
telnyx-agent setup-voice       # Create SIP connection + buy number + assign
telnyx-agent setup-ai          # Create AI assistant + buy number + wire together
telnyx-agent setup-porting     # Check portability + create porting order + submit
telnyx-agent status            # Account health overview

Every command supports --json for machine-readable output.

See Agent CLI

Model Context Protocol (MCP)

Telnyx hosts a remote MCP server at https://api.telnyx.com/v2/mcp.

To run a local Telnyx MCP server using npx:

npx -y @telnyx/mcp --api-key=YOUR_TELNYX_API_KEY

See MCP for more details about the generic API MCP proxy.

MCP Apps

tools/mcp-apps contains app-layer MCP servers with MCP Apps UI resources for focused Telnyx workflows. These are separate from the generic @telnyx/mcp proxy above.

Current apps:

• Number Intelligence (tools/mcp-apps/apps/number-intelligence)
• Usage & Cost Explorer (tools/mcp-apps/apps/usage-cost-explorer)
• Voice Monitor (tools/mcp-apps/apps/voice-monitor)

From tools/mcp-apps, use npm install, npm run typecheck, npm run build, and npm test.

Guides

Curl-first operational guides for common Telnyx workflows — SMS messaging, voice call control, AI assistants, phone numbers, porting, verification, webhooks, 10DLC registration, WireGuard networking, x402 payments, and Edge Compute handoff patterns.

For Edge Compute specifically, the goal is to make the handoff testable fast: start from a real telnyx-edge example, deploy it, and let team-telnyx/ai orchestrate against that live endpoint.

See Guides for the full list.

Edge Compute

team-telnyx/ai does not currently own native Edge Compute lifecycle support.

Instead, this repo should be treated as the orchestration/discoverability layer, while the actual function lifecycle lives in the separate team-telnyx/edge-compute repo and the telnyx-edge CLI.

In practice:

• use team-telnyx/ai for agent workflows, capability discovery, and AI-oriented integration patterns
• use team-telnyx/edge-compute + telnyx-edge for function creation, deployment, secrets, bindings, and lifecycle management

The intended end state is a clean bridge:

• ai = orchestrates and explains
• Edge Compute = deploys and runs (prefer API-key auth for agent flows)
• the boundary between them is a documented HTTP/MCP/function contract

See Edge Compute guide.

License

MIT