ElevenLabs Agents
by ElevenLabs
Platform for building real-time voice and chat AI agents
Last reviewed 2026-06-19
ElevenLabs Agents is a platform for building, deploying, and operating real-time AI agents that talk, type, and take action across voice and chat. It pairs ElevenLabs' own text-to-speech (with voice cloning and dozens of languages) and its Scribe speech-to-text with an orchestration layer that handles turn-taking, interruptions, and sub-second latency. Agents are built no-code in a web builder or programmatically via SDKs and a WebSocket API, then deployed to phone (telephony), web widgets, chat, and apps. The reasoning layer is LLM-agnostic: builders pick Claude, GPT, or Gemini, or bring a custom LLM. Agents connect to knowledge bases for retrieval, call external systems via server tools and function calling and MCP servers, and run omnichannel across phone, chat, and web. Telephony is first-class, with native Twilio plus other carriers and generic SIP. The product targets developers building voice features and mid-market and enterprise teams deploying support, scheduling, and sales agents. ElevenLabs differentiates on audio quality and latency; the reasoning lives in a third-party LLM.
What it can do
Answer and route phone calls via telephony
SupervisedHandles inbound and outbound calls over native Twilio and SIP with sub-second latency and turn-taking.
sourceCall external systems and execute actions
SupervisedInvokes server tools, function calls, and MCP servers to take actions in connected systems within configured guardrails.
sourceRetrieve and answer from a knowledge base
CopilotPulls from a connected knowledge base to answer questions during a live conversation.
sourceAdd voice to an existing chat agent
AssistantProvides STT and TTS so a developer's own LLM-driven chat agent can speak and listen.
source
Strengths
- +LLM-agnostic on top of best-in-class proprietary text-to-speech and its own speech-to-text
- +Strong first-class telephony story (Twilio, SIP, multiple carriers) with sub-second latency
- +Flexible build paths: no-code builder, SDKs, WebSocket API, function calling, and MCP
Limitations
- −Layered, recently changed pricing (bundled minutes plus overage plus pass-through LLM cost) makes total cost hard to predict
- −Autonomy is bounded to configured tools, not autonomous out of the box
- −Differentiation is audio quality, not reasoning, which lives in a third-party LLM
Overview
ElevenLabs Agents is a platform for building and running real-time voice and chat agents. It combines ElevenLabs' proprietary TTS and Scribe STT with an orchestration layer for turn-taking and low latency, and lets you pick the LLM.
What it does
Agents answer and place phone calls over telephony, call external systems via tools and MCP, retrieve from knowledge bases, and can add voice to an existing chat agent. Autonomy is bounded by the tools and guardrails you configure, so in practice it is a supervised agent.
Integrations & setup
Build no-code in the web builder or with SDKs (JS, React, Python) and a WebSocket API. Telephony via Twilio, Genesys, Vonage, Telnyx, Plivo, and SIP; tool integrations via function calling and MCP; CRM and scheduling connectors.
Pricing
Subscription plans plus usage: a free tier and paid tiers, with per-minute agent overage and separately billed LLM token costs. Verify current rates on the pricing page.
Best for / not for
Best for developers and teams that need high-quality, low-latency voice agents with their own choice of LLM. Less ideal where predictable flat pricing or out-of-the-box autonomy is the priority.
Alternatives
Vapi, Bland AI, Retell AI, and Synthflow are voice-agent platforms competing on latency, telephony, and build experience.
What people are saying
We aggregate real LinkedIn discussion into sentiment for the agents people search most. ElevenLabs Agents isn't tracked yet, want it added? Request tracking.
FAQ
Which LLMs can ElevenLabs Agents use?+
It is LLM-agnostic: you can select Claude, GPT, or Gemini, or bring a custom model. ElevenLabs supplies the text-to-speech and speech-to-text and the real-time orchestration.
Can ElevenLabs Agents make and take phone calls?+
Yes. Telephony is first-class via native Twilio and SIP, plus other carriers, for both inbound and outbound calls with low latency.
Sources
- ElevenLabs Agents · accessed 2026-06-19
- ElevenLabs Conversational AI · accessed 2026-06-19
- ElevenLabs pricing · accessed 2026-06-19
- ElevenLabs Series D · accessed 2026-06-19
Last reviewed 2026-06-19