LLM

Callplane + OpenAI Realtime

Speech-to-speech in a single model — when you need it.

OpenAI's Realtime API collapses STT + LLM + TTS into a single multimodal model — no transcription roundtrip, no TTS handoff. Callplane supports it as both a primary path (full speech-to-speech) and a hybrid path (Realtime LLM with separate Deepgram STT and a different TTS).

Why connect OpenAI Realtime to Callplane?

A voice AI assistant is only useful when it can connect to the rest of your stack. The OpenAI Realtime integration helps Callplane fit into the systems your team already uses for calling, transcription, reasoning, speech, routing, or customer data. That means the assistant can move from a demo into a production workflow without forcing you to replace your existing tools.

Callplane keeps the integration flexible. You can run one assistant with OpenAI Realtime, pair it with other providers for a full voice pipeline, and change providers later if your cost, latency, coverage, or quality requirements change. This is especially useful for teams that need different settings per assistant, region, campaign, or customer segment.

The result is a voice agent that is easier to operate: fewer hard-coded assumptions, clearer provider boundaries, and a pricing model that separates Callplane's platform fee from the third-party services you already trust.

What you get with Callplane + OpenAI Realtime

Native speech-to-speech with the lowest theoretical latency floor
Mix-and-match: use OpenAI Realtime for LLM only, separate STT/TTS
Function calling, tool use, and structured outputs
Multilingual conversation support
Voice activity detection and turn detection built-in

When to choose OpenAI Realtime

Choose OpenAI Realtime when absolute lowest latency is the top priority and you're OK with OpenAI's voice library. Choose Anthropic Claude or Google Gemini via Callplane's standard pipeline when reasoning quality or longer context matters more than the latency win.

How to configure

# .env
OPENAI_API_KEY=sk-...

# Per-assistant config:
# Mode: speech-to-speech  # or "split" for STT + LLM + TTS
# LLM -> Provider: OpenAI Realtime
# Model: gpt-4o-realtime-preview
# Voice: alloy   # or coral, echo, etc.

For full setup, see the docs.

Pricing

Callplane: $0.03/min platform fee. OpenAI Realtime: ~$0.06/min audio input, ~$0.24/min audio output on gpt-4o-realtime. Premium tier — best for high-value calls where latency and quality both matter. BYO mode passes OpenAI cost through directly.

Pairs well with

Telephony

Twilio

The default Callplane telephony stack — proven, global, programmable.

Telephony

Telnyx

Carrier-grade telephony at lower per-minute cost than Twilio.

STT

Deepgram

The fastest, most accurate STT for production voice agents.

Ship a voice agent on OpenAI Realtime today

200 free minutes. No credit card. Five-minute setup.

Start Free