> Open-source SDK for voice AI. Connect any AI agent to phone calls in 4 lines of code — pip install getpatter or npm install getpatter. Python & TypeScript, MIT licensed, sub-500ms latency. Canonical URL: https://www.getpatter.com/ Site map: https://www.getpatter.com/sitemap.xml · Overview for LLMs: https://www.getpatter.com/llms.txt · Developer resources: https://www.getpatter.com/developers --- Open-source · MIT # Give your AI agent a phone number The open-source SDK for voice AI. Connect any agent to real phone calls in 4 lines of code. Python and TypeScript, MIT licensed, sub-500 ms latency. Or call our agent yourself — [+1 (920) 7-PATTER](tel:+19207728837) US number · standard carrier rates apply main.py Live call · 00:14 ```python from getpatter import Patter, Twilio, OpenAIRealtime # Bind a Twilio number to a Patter instance phone = Patter(carrier=Twilio(), phone_number="+15550001234") # Wire up the agent with prompt + engine agent = phone.agent(engine=OpenAIRealtime(), system_prompt="Be a friendly receptionist.") # Spawn a Cloudflare tunnel and start handling calls await phone.serve(agent, tunnel=True) ``` 📞 +1 (415) 555-0142 · San Francisco, CA caller Hi, I'd like to book an appointment for next Tuesday. agent Tuesday's wide open. Morning or afternoon? caller Late morning if possible. agent Got it — 11:15 it is. TTFA · 412 ms STT · Deepgram TTS · ElevenLabs Built to slot into Twilio Telnyx OpenAI Realtime ElevenLabs Deepgram Anthropic Cartesia Twilio Telnyx OpenAI Realtime ElevenLabs Deepgram Anthropic Cartesia Features ## Voice AI in 4 lines of code. Connect your agent to real phone calls without the telephony headache. Compare what you'd hand-write against what Patter does for you. Without Patter ### Roll your own pipeline 371 lines ```python import asyncio, websockets, json from twilio.rest import Client from twilio.twiml.voice_response import VoiceResponse from fastapi import FastAPI, WebSocket import deepgram, httpx, audioop app = FastAPI() twilio = Client("SID", "TOKEN") # Start outbound call with TwiML call = twilio.calls.create( to="+1...", from_="+1...", url="https://your-server.com/twiml", ) # Serve TwiML to connect media stream @app.route("/twiml") async def twiml(): resp = VoiceResponse() connect = resp.connect() connect.stream(url="wss://your-server.com/ws") return str(resp) # Handle WebSocket audio stream @app.websocket("/ws") async def ws_handler(ws: WebSocket): await ws.accept() dg = deepgram.DeepgramClient("DG_KEY") dg_ws = dg.listen.live.v("1") # Transcode mulaw 8kHz → PCM 16kHz # Process STT transcripts # Send text to OpenAI / your LLM # Stream response to ElevenLabs TTS # Resample 24kHz → 8kHz mulaw # Stream audio back over WebSocket # Handle barge-in, call state, errors... # ... 60+ more lines of boilerplate ``` main.py · partial + webhook signing, retry, recording… With Patter SDK ### Just the agent. 4 lines app.ts ```typescript import { Patter, Twilio, OpenAIRealtime } from "getpatter"; // Bind a Twilio number to a Patter instance const phone = new Patter({ carrier: new Twilio(), phoneNumber: "+15550001234" }); // Wire up the agent with prompt + engine const agent = phone.agent({ engine: new OpenAIRealtime(), systemPrompt: "Be a friendly receptionist." }); // Spawn a Cloudflare tunnel and start handling calls await phone.serve({ agent, tunnel: true }); ``` app.ts 367 lines saved Build ## What you can build. A small, fast, opinionated set of primitives. Inbound, outbound, end-to-end voice, pipeline mode — same handler. 01 / Inbound & Outbound #### One handler, both directions. Receive incoming calls and place outbound dials with the same agent code. Native scheduling, voicemail detection, DTMF. ↘ Inbound ↗ Outbound 02 / Latency #### <500 ms time-to-first-audio. Native 16 kHz streaming, frame-level buffering, first-flush optimization. Sub-500 ms TTFA on Realtime mode. 0 ms 412 ms 1 s 03 / Architectures #### Two voice modes. End-to-end engines (OpenAI Realtime, ElevenLabs ConvAI) for the lowest latency, or a pluggable pipeline with 6 STTs, 5 LLMs, 5 TTSs. OpenAI Realtime end-to-end Pipeline · DG → 4o → EL composable 04 / Tools #### Function calls, on the phone. Point Patter at any function that returns a string. Agents can search a database, hit your API, or transfer to a human, mid-call. `@phone.tool` async def book_appt(date: str): … 05 / Guardrails #### Barge-in, transfer, recording. Production safeguards out of the box. Interruptions land cleanly. Transfer to a human warm or cold. Record both sides to disk or S3. barge-in transfer record DTMF 06 / Dev loop #### Tunnel, test, ship. `tunnel: true` spawns a Cloudflare tunnel and points your number at it. Terminal test mode lets you simulate a call from your shell. ▸ patter dev → tunnel: https://abc.trycloudflare.com → ready · listening for calls Inspect ## Monitor every call in real time. A built-in dashboard streams call metrics, transcripts, and per-segment cost — STT, LLM, TTS, audio — without leaving your machine. 01 ##### Live metrics Total calls, costs, latency, and active calls — streamed live. 02 ##### Call history Browse every call with full details, transcript, and per-stage cost. 03 ##### Export & filter Export to CSV or JSON with date filters for reporting and evals. 04 ##### Per-segment cost STT, LLM, TTS, carrier — see exactly what each call cost you. Patter dashboard · v0.5.5 3 live · 12 today +1 555 000 1234 ### Calls Real-time view of every call routed through this Patter instance. 1h 24h 7d All Total calls 128 ↑ 14% vs yesterday Avg latency p95 512 ms ↑ 38 ms vs yesterday Spend (24h) $8.42 saved $1.20 from cache Active now 3 2 inbound · 1 outbound #### Recent calls streaming · SSE | Status | From → To | Carrier | Duration | P95 latency | Cost | |---|---|---|---|---|---| | live | +1 415 555 0182 → +1 555 000 1234 | Twilio | 01:42 | 488ms | $0.06 | | live | +44 20 7555 0144 → +1 555 000 1234 | Telnyx | 00:37 | 372ms | $0.02 | | live | +1 555 000 1234 → +1 212 555 0167 | Twilio | 02:18 | 744ms | $0.09 | | ended | +1 718 555 0123 → +1 555 000 1234 | Twilio | 04:02 | 540ms | $0.18 | | ended | +1 305 555 0188 → +1 555 000 1234 | Telnyx | 01:23 | 418ms | $0.04 | | no answer | +1 555 000 1234 → +1 415 555 0173 | Twilio | 00:00 | — | $0.00 | | ended | +1 408 555 0199 → +1 555 000 1234 | Twilio | 03:11 | 502ms | $0.13 | ##### Latency · last 24h P50 total 412 ms P95 total 512 ms STT avg 122 ms LLM avg 218 ms TTS avg 128 ms Network 42 ms Ship ## Build your voice agent on top of Patter. ##### You build the agent — we connect the phones. Telephony, audio streaming, and carrier routing are taken care of. You write the prompt, the tools, and the logic. ##### Any LLM, any voice. OpenAI, Anthropic, Groq, Cerebras, Google for LLM. 6 STTs and 5 TTSs. Provider-agnostic — startups and enterprises ship faster without lock-in. ##### Production ready from day one. Call recording, transfer, DTMF, voicemail detection, barge-in, automatic retries — all built in, zero extra infrastructure. The Patter Stack 3 layers Customize ## Fully customizable based on your needs. Every layer of the stack is pluggable. Match latency, quality, cost, and compliance — without vendor lock-in. 27+ provider integrations across the voice stack 3 voice modes — Realtime, Pipeline, Hybrid 2 SDKs — Python & TypeScript at parity LLM text generation ##### Pick a model, swap it tomorrow. OpenAI Anthropic Google Gemini Groq Cerebras STT speech-to-text ##### Six transcribers, one interface. Deepgram AssemblyAI Cartesia Soniox Speechmatics Whisper TTS text-to-speech ##### Voices, hot-swappable mid-call. ElevenLabs OpenAI Cartesia LMNT Rime Telnyx Realtime all-in-one voice ##### End-to-end engines. OpenAI Realtime Gemini Live Ultravox ElevenLabs ConvAI Telephony phone carriers ##### BYO carrier, BYO number. Twilio Telnyx · beta Audio VAD & suppression ##### Cleaner audio, sharper turns. Silero VAD Krisp DeepFilterNet - LLM fallback chain. Automatic provider failover, mid-call. - Terminal test mode. Simulate a call from your shell. - Built-in Cloudflare tunnel. Public webhook in 3 seconds. - Pipeline hooks. Middleware on STT, LLM, and TTS stages. - Scheduled outbound. Cron and one-off scheduling, from the SDK. - Eval harness. Regression testing with LLM-as-judge. - Built-in dashboard. Real-time metrics, history, CSV/JSON export. - 1,766 tests. Run them yourself; no telemetry, no lock-in. Offering ## 100% open source, self-hosted. Run Patter on your own infrastructure with your own provider keys. Full control, no lock-in — free under the MIT license. 40+ features · growing every release ### From `npm install` to first call: ~5 minutes. Permissive MIT license. No usage caps. No telemetry. The same SDK that runs in our examples runs in your production. [View on GitHub →](https://github.com/PatterAI/Patter) [Read the docs →](https://docs.getpatter.com) License MIT Tests 1,766 ✓ Stars ★ Releases weekly README · top of file ``` $ npm install getpatter # or: pip install getpatter $ export TWILIO_ACCOUNT_SID=AC... $ export TWILIO_AUTH_TOKEN=... $ export OPENAI_API_KEY=sk-... $ npx patter dev app.ts # tunnel up · https://abc.trycloudflare.com # ready · listening for calls on +1 555 000 1234 $ npx patter call --to +15558675309 → dialing... ringing... connected. → TTFA: 386ms · STT: deepgram · TTS: elevenlabs ``` Compare ## No marketing math. An open-source SDK between hosted platforms and self-hosted frameworks. Side-by-side with Vapi, Retell, LiveKit Agents, and Pipecat — what Patter does, what it doesn't, what you're choosing. | | Patter | Vapi | Retell | LiveKit Agents | Pipecat | |---|---|---|---|---|---| | Architecture | Embedded SDK | Hosted platform | Hosted platform | SDK + server | Framework | | Self-hosted, zero vendor infra | ✓ | — | — | ✓ | ✓ | | TypeScript + Python parity | ✓ | ~ | ~ | ✓ | — | | Per-segment cost dashboard | ✓ | ~ | ~ | — | — | | Built-in Cloudflare tunnel | ✓ | — | — | — | — | | End-to-end Realtime engines | ✓ | ✓ | ~ | ✓ | ✓ | | License | MIT | Closed platform | Closed platform | Apache 2.0 | BSD-2 | ✓ Fully supported ~ Partial · behind hosted dashboard — Not supported What you can build ## Any voice workflow, powered by your agent. Start building with Patter and see what's possible. The starter templates ship as standalone repos. 01 / Support #### Customer support, 24/7. AI phone agents handle tier-1 support around the clock. Route the messy ones to humans, mid-call. ↓ 70% answered < 30 sec 02 / Reminders #### Appointment reminders. Outbound calls confirm, reschedule, or cancel. Natural conversation, not a robot script. ↓ 60%+ no-shows 03 / Sales #### Lead qualification. Agents call leads, qualify with natural conversation, sync to your CRM. Same handler, both directions. ↑ 3x SDR throughput 04 / Research #### Surveys & feedback. Run NPS and customer surveys at scale through natural voice. Transcripts ready for analysis. ↑ 4x response rate ## Try it. Then ship it. Call our agent. Read the four-line quickstart. Star the repo. Or hop on a 20-minute call with us — we'll help you wire your first agent. [Get a demo →](https://www.getpatter.com/book-call) [View on GitHub](https://github.com/PatterAI/Patter) [Read the docs](https://docs.getpatter.com) Or call us right now +1 (920) 7-PATTER US number · standard carrier rates apply · routes to a Patter agent running on this exact SDK --- Engine OpenAI Realtime 2 TTFA 412 ms