Voice Agents Python Packages

calibrate-sdk

Python SDK for Calibrate's API

pipecat-mirai

Mirai for Pipecat: Hindi, Hinglish and Gujarati text-to-speech, the Realtime API (speech to speech), and smooth phone-call audio

premove-itn

Premove ITN: context-aware inverse text normalization for English voice-agent transcripts.

dunetrace

The runtime reliability layer for AI agents.

tring

Open-source voice agent stack: one agent contract, any runtime (cascade / speech-to-speech / hybrid), local-first, cost-transparent.

hotato

Find what broke in your agent calls. Pin it so it never ships again. Local voice-agent call forensics and regression guards.

streamdouble

A framework-agnostic Twilio Media Streams simulator. Test your voice agent's WebSocket endpoint locally, without placing a real call.

onepin

Onepin Python SDK + CLI — orchestrate, validate & ship production-ready voice audio across 100+ TTS models

yapitalism

Drive terminal coding agents by voice, with receipts that never claim more than they proved

call2me

Official Call2Me Python SDK — Build AI voice agents in minutes

unpod

SDK to use unpod's calling infra and speech pipeline

spekoai

Official Speko Python SDK (spekoai) - one API for every voice AI provider: STT, TTS, and LLM routing + failover, sync + async

attenlabs-saa

Addressee detection for voice agents: device-directed speech detection that runs before STT, so background speech, side conversations, and the agent's own TTS echo never trigger it. No wake word, model-agnostic, drop-in for LiveKit, Pipecat, ElevenLabs, Twilio, and OpenAI. The layer your VAD and turn detection are missing.

callback-voice

Chaos testing for voice agents: simulated callers, real audio, one exit code.

voice-evals

Reproducible evaluation for voice agents: replay scoring, scripted live probes over WebSocket, and CI gates for WER, latency, barge-in behavior and task outcomes.

indicorderbench

Catches multilingual voice agents placing the wrong order. τ-bench style benchmark + CI gate for voice ordering agents in English and Hinglish: checks the committed order, not the transcript.

pipecat-bithuman

Moved to GitLab: https://gitlab.com/bithuman/sdk/pipecat-bithuman (this GitHub copy is archived, read-only)

makimoto-kawa

Official Python SDK for Makimoto. Build speech-to-text, real-time transcription, call summarisation, tagging, and conversational AI applications with a simple developer-friendly API.

saa-pipecat-client

Addressee detection for voice agents: device-directed speech detection that runs before STT, so background speech, side conversations, and the agent's own TTS echo never trigger it. No wake word, model-agnostic, drop-in for LiveKit, Pipecat, ElevenLabs, Twilio, and OpenAI. The layer your VAD and turn detection are missing.

saa-livekit-client

Addressee detection for voice agents: device-directed speech detection that runs before STT, so background speech, side conversations, and the agent's own TTS echo never trigger it. No wake word, model-agnostic, drop-in for LiveKit, Pipecat, ElevenLabs, Twilio, and OpenAI. The layer your VAD and turn detection are missing.

outturn

Did the agent produce the right outcome? As a number.

dograh-sdk

Open source voice AI platform. Self-hosted alternative to Vapi and Retell. On Prem, BYOK across Speech to Speech or LLM/STT/TTS, with a visual workflow builder, MCP native and telephony support.

livekit-plugins-voicemem

Long-term memory for LiveKit voice agents, backed by PostgreSQL and pgvector. Remembers both what the caller said and what they are like.

voxprobe

Persona-driven QA for voice agents: adaptive simulated callers, in-process Pipecat audio arena, both-sides recording, audio-timed transcripts, turn-taking/latency metrics, evidence-backed findings