Patter

Open-source voice-AI SDK. The Vapi/Retell alternative for builders who want to own the stack. Give your AI agent a phone number in 4 lines — Python and TypeScri

What is Patter?

Open-source voice-AI SDK. The Vapi/Retell alternative for builders who want to own the stack. Give your AI agent a phone number in 4 lines — Python and TypeScript, MIT licensed, Twilio, Telnyx, and Plivo.

Overview of Patter

What it is

Patter is an open-source SDK that connects AI agents to the phone network. It handles the agent loop, language model, speech-to-text, text-to-speech, real-time voice, audio processing, and telephony carrier integration. It is offered under the MIT license and supports Python and TypeScript. The SDK is designed as an alternative to proprietary voice-AI services like Vapi and Retell, giving developers full control over the stack.

Key features

Supports 27+ provider integrations across LLM, STT, TTS, realtime, telephony, and audio layers.
Choose from providers such as OpenAI, Anthropic, Google Gemini, Groq, Cerebras, Deepgram, AssemblyAI, ElevenLabs, Twilio, Telnyx, and Plivo.
Three voice modes: Realtime, Pipeline, and Hybrid.
Built-in LLM fallback chain for provider failover mid-call.
Vendor-neutral OpenTelemetry trace for each call.
Skills bundle for AI coding agents that works with ~55 agent harnesses.

Use cases

Answering phone calls as a receptionist or booking assistant.
Making outbound calls with answering machine detection and voicemail drop.
Integrating tool calling (e.g., CRM lookup, ticket creation) via webhooks.
Building custom voice pipelines with specific STT and TTS providers.
Running a production-ready setup with tools, guardrails, recording, and a dashboard.

Getting started

bash# TypeScriptnpm install getpatter# Pythonpip install getpatterSee the repository README for full quickstart code examples in TypeScript and Python.

License and pricing

Licensed under MIT. Pricing is not stated; the repository is marked as open source.---

這是什麼

Patter 是一個開源 SDK，能將 AI 代理連接到電話網路。它處理代理循環、語言模型、語音轉文字、文字轉語音、即時語音、音訊處理以及電信營運商整合。採用 MIT 授權，支援 Python 和 TypeScript。此 SDK 被設計為 Vapi 和 Retell 等專有語音 AI 服務的替代方案，讓開發者完全擁有堆疊的控制權。

主要功能

支援超過 27 個提供者整合，涵蓋 LLM、STT、TTS、即時通訊、電信和音訊層。
可選擇 OpenAI、Anthropic、Google Gemini、Groq、Cerebras、Deepgram、AssemblyAI、ElevenLabs、Twilio、Telnyx 和 Plivo 等提供者。
三種語音模式：即時、管線和混合。
內建 LLM 備援鏈，可在通話中進行提供者故障轉移。
每通電話的供應商中立 OpenTelemetry 追蹤。
為 AI 編碼代理設計的技能套件，適用於約 55 種代理框架。

使用情境

接聽電話，作為接待員或預訂助手。
撥打外撥電話，具備答錄機檢測和語音信箱投遞功能。
透過 Webhook 整合工具呼叫（例如 CRM 查詢、建立工單）。
使用特定的 STT 和 TTS 提供者建立自訂語音管線。
運行配備工具、護欄、錄音和儀表板的生產環境設定。

快速開始

bash# TypeScriptnpm install getpatter# Pythonpip install getpatter請參閱儲存庫 README 以取得完整的 TypeScript 和 Python 快速入門程式碼範例。

授權與價格

採用 MIT 授權。價格未提及；該儲存庫標記為開源。}jsonjson```jsonbold textWe need to produce a JSON object with two keys

Based on the project's public README and repository metadata.