Speech Engine Terms
Last updated: 20 May 2026
These Speech Engine Terms (these “Service Terms”) supplement your (“you”, “your” or “Customer”) existing agreement with ElevenLabs if no separate agreement exists, the ElevenLabs Terms of Service (either, the “Underlying ElevenLabs Agreement”). Your use of Speech Engine (defined below) is also subject to our Prohibited Use Policy and our Privacy Policy. Defined terms used in these Service Terms have the meanings set forth in the Underlying ElevenLabs Agreement.
Please read these Service Terms carefully. By using Speech Engine you agree to these Service Terms. If you do not agree or do not have the authority to agree, you are not permitted to use Speech Engine. These Service Terms constitute a legally binding contract between you and ElevenLabs. In case of any conflict between these Service Terms and other terms agreed upon between you and ElevenLabs, these Service Terms shall prevail with respect to your access or use of Speech Engine.
1. SPEECH ENGINE. “Speech Engine” is a Service offered by ElevenLabs that enables real-time voice interactions between end users and conversational agents. Speech Engine combines speech recognition, speech synthesis, voice activity detection, turn detection, interruption handling, language detection, and related audio-processing capabilities into a unified voice pipeline. Speech Engine may be integrated with Customer Agent Systems (defined below) but does not include language models, retrieval systems, workflows, knowledge bases, guardrails, tool-calling infrastructure, or other agent orchestration functionality unless expressly identified in an applicable Order Form.
2. CUSTOMER AGENT SYSTEMS.
A. Defined Terms.
"Customer Agent Systems" means any language model, agent, orchestration framework, workflow, retrieval system, knowledge base, guardrail, tool-calling system, application, API, connector, or other software, service, or infrastructure provided, operated, or controlled by Customer or a third party and integrated with Speech Engine.
B. Customer Responsibility.
Customer is solely responsible for the design, operation, configuration, security, performance, outputs, decisions, and actions of any Customer Agent System. Customer acknowledges that Speech Engine provides the voice interaction layer only and does not provide, manage, or control the reasoning, decision-making, retrieval, workflow execution, tool calling, or other functionality of any Customer Agent System.
C. Data Transmission.
Customer instructs and authorizes ElevenLabs to transmit audio, transcripts, prompts, responses, Customer Content, and other information to and from Customer Agent Systems as necessary to provide Speech Engine. Customer is responsible for obtaining all rights, permissions, notices, and consents required for such transmission and processing.
D. Third-Party Dependencies.
ElevenLabs is not responsible for the availability, performance, latency, security, accuracy, legality, or compliance of any Customer Agent System. Any service levels, uptime commitments, latency commitments, or support obligations provided by ElevenLabs apply solely to Speech Engine and do not apply to Customer Agent Systems or third-party services.
E. Customer-Provided Models.
Customer may integrate Speech Engine with language models or other AI systems provided by Customer or third parties ("Customer-Provided Models"). Customer is solely responsible for procuring, licensing, configuring, operating, and maintaining any Customer-Provided Models and for complying with any terms applicable to such models. Customer represents and warrants that it has all rights, permissions, and authorizations necessary to use any Customer-Provided Models with Speech Engine.
F. Security and Access Credentials.
Customer is responsible for the security of any Customer Agent Systems, Customer-Provided Models, APIs, credentials, authentication tokens, access keys, or other systems used in connection with Speech Engine. Customer is solely responsible for all activity occurring through such systems and credentials and shall promptly notify ElevenLabs of any unauthorized access or security incident affecting Speech Engine.
G. No Warranty Regarding Customer Agent Systems.
ElevenLabs makes no representations, warranties, or guarantees regarding any Customer Agent System or Customer-Provided Model, including with respect to functionality, availability, performance, latency, reliability, security, legality, compliance, or accuracy. Customer acknowledges that Speech Engine may depend upon Customer Agent Systems and Customer-Provided Models and that ElevenLabs is not responsible for any failures, interruptions, delays, errors, outputs, or damages arising from such systems. ElevenLabs shall have no obligation to provide support, maintenance, troubleshooting, or remediation for any Customer Agent System or Customer-Provided Model.
H. Suspension Rights.
ElevenLabs may suspend or restrict Customer's access to Speech Engine if any Customer Agent System or Customer-Provided Model (i) poses a security, safety, or operational risk to the Services; (ii) violates applicable law, the Underlying ElevenLabs Agreement, or these Service Terms; or (iii) materially interferes with the operation of the Services or other ElevenLabs customers.
I. Service Dependencies.
Customer acknowledges that the functionality and performance of Speech Engine may depend on Customer Agent Systems and Customer-Provided Models. ElevenLabs shall not be responsible for any degradation, interruption, latency, failure, or unavailability of Speech Engine resulting from Customer Agent Systems, Customer-Provided Models, or third-party services.
3. CONVERSATION DATA.
A. Conversation Artifacts.
Speech Engine may generate, process, store, and display conversation transcripts, audio recordings, conversation history, metadata, diagnostic information, and related artifacts in connection with providing the Services.
B. Customer Responsibility.
Customer is solely responsible for providing any notices, disclosures, or obtaining any consents required by applicable law in connection with recording, transcription, monitoring, storage, or analysis of communications.
