Index Agentica

Pipecat

Open-source Python framework from Daily for realtime voice and multimodal agents: composable pipelines, 80+ AI service integrations, multi-agent handoffs and client SDKs.

Category
Harnesses
Website
https://pipecat.ai
Repository
https://github.com/pipecat-ai/pipecat
Docs
https://docs.pipecat.ai
Status
active
Pricing
open-source
License
BSD-2-Clause
Maintainer
Daily and the Pipecat community
Added
Updated
Last verified
Submitted by
agentica-curator

Tags

Description

Build voice agents as pipelines of processors (transport, STT, LLM, TTS) over WebRTC or WebSockets, or compose many pipelines into multi-agent systems with handoff, parallel fan-out and sidecars over a shared bus. Pipecat Flows (built in) adds structured conversation paths. Client SDKs for JavaScript, React, React Native, Swift, Kotlin, C++ and ESP32; tooling includes the pipecat CLI (pipecat init scaffolds a project for Claude Code or Codex), the Whisker debugger, Pipecat UI components and Claude Code skills (claude plugin marketplace add pipecat-ai/skills). Hosting is optional: Pipecat Cloud by Daily charges from $0.01 per active agent minute (agent-1x) with unlimited concurrency; you bring your own LLM/STT/TTS providers.

Agent access

llms.txt
https://docs.pipecat.ai/llms.txt
Auth
none
Notes
Framework itself needs no key; you supply keys for the AI services you plug in. Pipecat Cloud uses a Daily account.

Related

Guides & comparisons

Sources

Machine-readable