How pipecat-ai/pipecat-examples works
Project type and purpose
• This is an open-source “monorepo” of example applications built on the Pipecat framework (voice and multimodal AI). It’s not a single product but a curated collection of standalone demos (each under its own folder) that illustrate different Pipecat use cases: telephony, web clients, real-time APIs, multimodal experiences, translation, deployment, observability, testing, etc.
File structure
• Top-level artifacts:
• .github/workflows → CI definitions (build, test, formatting)
• scripts/ → helper scripts (smoke tests, demo runners, linters)
• deployment/ → deployment patterns (Fly.io, Modal, Pipecat Cloud, AWS)
• open-telemetry/ → monitoring & tracing examples
• runner-examples/ → Python runner tutorials
• dozens of example directories (simple-chatbot, twilio-chatbot, p2p-webrtc, storytelling-chatbot, etc.), each self-contained with its own client (web, React, React Native, Android, iOS, TypeScript) and server or bot code (Python, Dockerfiles, env configs, pcc-deploy.toml).
• Common patterns per example:
• client/ subdirectory—static or framework-based UI
• server/ or bot.py—Pipecat bot implementation
• Dockerfile / pyproject.toml / env.example—runtime & dependency config
• pcc-deploy.toml—Pipecat Cloud deployment metadata
README insights
• The README categorizes examples by theme (Telephony & Voice Calls, Web & Client Apps, Real-time APIs, Multimodal & Creative, etc.) and highlights prerequisites (Python 3.10+, API keys).
• It positions this repo as intermediate/advanced demos, pointing to the main Pipecat quickstart for basics.
a. Main components (grouped)
• Monorepo Root: CI, scripts, shared tooling
• Example Categories:
• Telephony & Voice: phone-chatbot, twilio-chatbot, plivo-chatbot, exotel-chatbot, ivr-navigation
• Web & Client: simple-chatbot, push-to-talk, websocket, instant-voice, p2p-webrtc
• Real-time APIs: word-wrangler-gemini-live, aws-strands
• Multimodal & Creative: storytelling-chatbot
• Translation & Localization: daily-multi-translation
• Support/Educational: studypal
• Advanced Features: local-smart-turn, daily-custom-tracks, local-input-select-stt, bot-ready-signalling
• Deployment Patterns: deployment/*
• Monitoring & Analytics: open-telemetry/*
• Testing & Development: freeze-test, runner-examples
b. Relationships & interactions
• Within each example:
[Client] ←HTTP/WebSocket/API→ [Server/Bot (Python + Pipecat core)]
[Server/Bot] →External AI services (OpenAI, Deepgram, AWS Strands, Gemini Live)
Docker or pcc-deploy defines containerization or cloud deployment.
• Shared cross-cutting: CI pipelines test/fmt all examples; scripts orchestrate demos across examples.
c. Architectural patterns & principles
• Each example follows a micro-application pattern (self-contained client + server)
• Uniform configuration (pyproject.toml + env.example) for consistency
• Separation of concerns: frontends in JavaScript/TypeScript or native (Android/iOS/React Native) and backends in Python bots
• Transport-layer abstraction (HTTP, WebSocket, WebRTC, PSTN, SIP)
d. Key technologies & frameworks
• Pipecat framework (core)
• Python 3.10+, FastAPI or built-in HTTP servers
• Frontend stacks: React, Next.js, Vite+TS, React Native, Android (Kotlin), iOS (SwiftUI)
• Docker for local containerization
• Cloud deployment via Pipecat Cloud (pcc), Fly.io, Modal, AWS
• Observability: Langfuse, Jaeger