Voice APIs and runtimes give a product team control over audio streaming, models, tools, state, and deployment. That control shifts more responsibility to your team: carrier setup, latency budgets, monitoring, retries, safety, and maintenance. Compare the runtime surface with the engineering work it leaves to you.
Questions to ask
Which audio, telephony, and SDK surfaces are documented?
How are interruption and retries measured?
What observability and data controls are included?
Programmable Voice component that connects a TwiML ConversationRelay session to an application WebSocket for STT/TTS, language settings, DTMF, and call events.
Telephony voice-agent runtime
Scope and signals
Best fit: Developers building their own voice-agent logic while keeping telephony, media handling, and call control in Twilio.
Voice stack: Cascaded speech interface with developer-supplied LLM
Google's Gemini Live API streams multimodal sessions over WebSockets; its stable native voice-agent models are Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking.
Native realtime multimodal model API
Scope and signals
Best fit: Developers building low-latency conversational agents that need native audio, multimodal input, and function calling in a persistent WebSocket session.
Voice stack: Native audio agents: Gemini 3.8 Live (gemini-3.8-live) and Gemini 3.8 Live Extended Thinking (gemini-3.8-live-extended-thinking)., Separate TTS models: gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts generate speech; neither is a Live agent., Separate transcription endpoints: gemini-3.5-transcribe handles audio files; gemini-3.5-transcribe-live is a Live API speech-to-text mode, not a conversational agent., Separate translation mode: gemini-3.5-live-translate-preview is a Live API interpreter for 70+ languages, not a conversational agent; it is preview.
Standard paid rates are $0.005/min equivalent for input audio and $0.018/min equivalent for output audio, billed separately plus text tokens; Google also lists a free API tier.
Hosted voice and chat agent platform built around either node-based conversation flows or single prompts, with telephony, tools, knowledge base, and analytics.
Voice agent platform
Scope and signals
Best fit: Teams comparing a managed agent builder with both visual flow design and API/SDK control.
Voice stack: Configurable STT-LLM-TTS pipeline with custom LLM option
Unified Voice Agent API that combines Deepgram speech recognition, LLM orchestration, and voice synthesis with function calling and mid-session controls.
Voice agent API and speech infrastructure
Scope and signals
Best fit: Developers who want one managed voice API instead of assembling separate STT, orchestration, and TTS services.
Voice stack: Cascaded STT-LLM-TTS managed full stack
What work can a voice AI platform leave to my team?
Expect to own some combination of carrier setup, realtime transport, state, prompts, tools, identity, retries, monitoring, evaluation, data policy, scaling, incident response, and human escalation. Confirm the boundary in writing.
Which layer of the voice application are you buying?
A speech API, an orchestration runtime and a managed voice platform sit at different layers. Write down who supplies audio transport, speech or realtime models, tool execution, telephony, transfer and monitoring before comparing vendors.
Use a platform when you need control over how an application handles a conversation and connects to your systems. For a packaged calling workflow, evaluate the phone-agent category first. Neither choice removes the need to test the complete call path.
Compare platforms with one integration exercise
Build the same short workflow: identify a caller, retrieve a record, confirm a requested change and transfer with context. Test a slow tool, interrupted speech and an unavailable destination.
Record which behavior you configured and which required code. Compare runtime logs, versioning, supported call surfaces and total operating cost, including engineering maintenance.
Evaluate language support in real calls
Test recognition, speech quality, interruptions, language switching, tool use, and human transfer with your own calls in the required language. Availability can vary by model, plan, and phone route.
DocsBot documents multilingual voice workflows and Japanese support. Verify the exact voice, phone path, and transfer behavior in your pilot. Review DocsBot language information ↗
Other providers’ language claims reflect cited vendor material; confirm Japanese and other required languages directly with each provider.
Listing does not imply endorsement or a performance guarantee. This directory is published by DocsBot. Editorial policy
Your business. Your shortlist.
Find your best-fit AI voice system.
Start with your website. Answer a few focused questions, then get a personalized PDF with fit, trade-offs, alternatives and a pilot plan for your call flow.
