ElevenAgents

Structured procedures compile on save and publish: Update agent and Create agent draft now compile and validate structured procedures automatically. Integrations no longer need to call the compile endpoint or manually update the generated workflow. Invalid procedures return per-procedure errors as 400 procedure_validation_failed. Agents without structured procedures are unaffected.

Compile procedures endpoint is legacy: Compile procedures is marked legacy and should not be used by new integrations. It remains available for existing callers.

Merge proposals: Merge proposals are available in alpha. Teams can open a proposal for an agent branch, review configuration changes and test results, leave comments or approval decisions, and merge approved changes into a target branch. See Merge proposals.

SDK releases

JavaScript SDK

v2.71.0 - Added clients for agent merge proposals, deployment history, test invocation cancellation and Speech Engine duplication. The release also adds triage ticket priorities and sorting, batch call export limits, gpt-6.1-sol, claude-sonnet-5-5, Eleven v4 agent voice models and current agent configuration types.

Python SDK

v2.71.0 - Added clients for agent merge proposals, deployment history, test invocation cancellation and Speech Engine duplication. The release also adds triage ticket priorities and sorting, batch call export limits, gpt-6.1-sol, claude-sonnet-5-5, Eleven v4 agent voice models and current agent configuration types.

Swift SDK

v3.4.0 - Dynamic variables now preserve native JSON types when sent in conversation initiation data.

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added realtime Speech to Text transcript editing through transcriptEdit, the edited_transcript event and React's onEditedTranscript callback. React now forwards filterBackgroundAudio to realtime sessions. The client also exports MessagePayload and uses stable response IDs to prevent duplicate agent replies.

@elevenlabs/[email protected] and @elevenlabs/[email protected] - Added image previews and download links for agent message attachments, including attachment-only replies. Stable response IDs prevent duplicate replies, and text-only conversations now use a stop control with an End chat tooltip.

API

Eleven v4 and Eleven v4 Turbo

Eleven v4 and Eleven v4 Turbo are now available. Eleven v4 provides higher-quality, more expressive speech and improved voice cloning across more than 90 languages. Eleven v4 Turbo provides the same model family for real-time applications, with median inference latency of approximately 100 ms.

Use eleven_v4 through the Text to Dialogue API for content creation and long-form audio. Use eleven_v4_turbo through the Text to Dialogue WebSocket for agents and interactive applications.

Speech to Text

Transcript editing: Batch and realtime transcription now accept a natural-language edit instruction of up to 2,000 characters. Batch responses return the result in edited_transcript, while realtime sessions emit an edited_transcript event. See the batch and realtime guides for incompatibilities, response handling and pricing.

ElevenAgents

Agent models: Added gpt-6-sol, gpt-6-luna, glm-52, deepseek-v41-flash, claude-opus-5 and claude-opus-5-5 to the supported agent LLM options.

Agent test usage: Test invocation summaries can include total credits and USD price. Individual test runs can include credits and a charging breakdown.

Flows

Template API: Added endpoints to list and retrieve published Flows templates, start template runs and retrieve run results. Runs can target a published version and send their terminal result to a configured webhook.

Text to Dialogue

Generation continuity: Text to Dialogue requests can use preceding or following text and request IDs to maintain continuity across generations.

Professional voice mode: Text to Dialogue requests can use the Instant Voice Clone version of a Professional Voice Clone to reduce latency or improve expressiveness.

SDK releases

JavaScript SDK

v2.70.0 - Added batch and realtime Speech to Text transcript editing types, branchId filtering and cost fields for test invocation summaries, knowledge base auto-discovery, Text to Dialogue usePvcAsIvc and the deepseek-v41-flash agent model.

Python SDK

v2.70.0 - Added batch and realtime Speech to Text transcript editing types, branch_id filtering and cost fields for test invocation summaries, knowledge base auto-discovery, Text to Dialogue use_pvc_as_ivc and the deepseek-v41-flash agent model.

ElevenLabs CLI

v1.4.0 - Fixed requests with an explicit API key so they no longer also send stored OAuth credentials, which caused authentication failures. Added descriptions for the feedback command group and instructions for reporting missing capabilities to generated agent skills. Regenerated the API client with Flows template endpoints and current schema types.

API

Image & Video

Sora 2 and Sora 2 Pro retired: OpenAI is discontinuing the Sora API on September 24, 2026. Both models have been removed from the Image & Video model picker and stop generating on that date. Existing generations remain available in your history. Flows and templates that use a Sora node need to be switched to another video model, such as Gemini Omni 1.1 Flash, before they can run again.

Seedance 1.5 Pro deprecated: ByteDance retires Seedance 1.5 Pro on November 11, 2026. The model is no longer offered for new generations and will stop working on that date. Use a Seedance 2.0 model instead.

ElevenAgents

Parallel tool calls: Agent prompt configuration now includes enable_parallel_tool_calls (boolean, default true). When enabled, supported models can execute multiple tools in one turn.

Agent models: Added gpt-6-astra to the agent LLM options.

Alerting: Agent alerting now supports Slack channels alongside PagerDuty and webhook notifiers.

Phone number search: Added a cursor-paginated phone number endpoint with filters for provider, outbound support, assigned agent and branch, label and phone number.

Image and video

GPT Image 2.5 models: Image generation now supports gpt-image-2.5-flare and gpt-image-2.5-sunburst. Both models accept up to 10 reference images, quality levels through max, 14 fixed aspect ratios plus auto, and 1K, 2K or 4K output.

Speech Engine

Cascade timeout: Speech Engine create and update requests now accept cascade_timeout_seconds (number, 2–15, default 4) to control how long ElevenLabs waits for an upstream speech engine before retrying.

Text to Speech

Multi-context turn boundaries: The is_final_audio_for_turn marker is now emitted after every buffered byte for that turn, including MP3 and Opus output. Clients can use the marker as an exact turn boundary without switching to PCM.

SDK releases

JavaScript SDK

v2.69.0 - Added transcriptEdit to the realtime Speech to Text wrapper, with edited text delivered through edited_transcript events. Fixed Speech Engine JWT verification for API keys with data residency suffixes. Added phoneNumbers.listV2 and Flows template methods to list and get templates and list, create and get runs. Regenerated clients and types for GPT Image 2.5, Slack agent alerting, parallel tool calls, gpt-6-astra, service account API key concurrency limits, Music zero retention, conversation filters and MCP tool approval statuses.

Python SDK

v2.69.0 - Added transcript_edit to the realtime Speech to Text wrapper, with edited text delivered through edited_transcript events. Fixed Speech Engine JWT verification for API keys with data residency suffixes, and made play raise an error when ffplay fails. Added phone_numbers.list_v2 and Flows template methods to list and get templates and list, create and get runs. Regenerated clients and types for GPT Image 2.5, Slack agent alerting, parallel tool calls, gpt-6-astra, service account API key concurrency limits, Music zero retention, conversation filters and MCP tool approval statuses.

ElevenLabs CLI

v1.3.2 - Added optional --intent metadata to collect feedback on CLI usage. Added elevenlabs feedback missing-capability to report workflows that the CLI does not support.

API

ElevenAgents

Call queueing and hold audio: Added call queueing for agents at their concurrency limit. Queued callers hear hold audio and receive queue_status events until they are admitted or time out. New API methods upload or delete custom hold audio.

Test version metadata: Agent test results now identify the agent version they ran against and whether they included draft changes. Invocation summaries also indicate when partial resubmissions ran against different versions.

Model and account metadata: Agent LLM configuration now accepts gemini-3.8-flash. WhatsApp account responses identify whether the account uses the Cloud API or coexistence signup flow.

Music

Music v2.5 API support: Music endpoints and SDKs now accept music_v2_5. Composition plan generation chunks support up to 6,132 characters, with up to 30 lines of 200 characters each.

SDK releases

JavaScript SDK

v2.68.0 - Added methods to upload and delete agent hold audio. Regenerated types for call queueing, procedure references, draft branch creation, RAG query limits, test version metadata, gemini-3.8-flash and music_v2_5.

Python SDK

v2.68.0 - Added synchronous and asynchronous methods to upload and delete agent hold audio. Regenerated types for call queueing, procedures, draft branch creation, RAG query limits, test version metadata, gemini-3.8-flash and music_v2_5. The compose_detailed wrapper now accepts every generated music model ID and both prompt- and chunk-based composition plans.

iOS SDK

v3.3.1 - Reconciles agent messages by response_id instead of event_id, preserving multiple responses around tool calls and preventing streamed parts or corrections from attaching to the wrong message.

ElevenLabs CLI

v1.3.0 - Added optional --intent metadata to generated API commands, with ELEVENLABS_AGENT_INTENT as a task-wide default. Added elevenlabs feedback missing-capability to report workflows that the CLI does not support.

Packages

@elevenlabs/[email protected] and @elevenlabs/[email protected] - Fixed duplicate agent replies after tool and procedure calls. Streamed text replies retain Markdown formatting, and voice conversations no longer render unspoken streamed chat parts. The embed package also limits user_activity events to once per second and clears the limit when a session ends.

API

ElevenAgents

Workspace-wide conversation tickets: Added List workspace tickets to retrieve conversation triage tickets across accessible agents. Filter by status or assignee and use cursor pagination to retrieve up to 100 tickets per page.

Dynamic variable conversation filters: List conversations and Text search conversation messages now accept repeatable dynamic_variable_params filters. Use name:op:value with eq, gt, gte, lt or lte; comparison operators require numeric values.

Twilio answering machine detection: Outbound telephony and batch call configurations now support optional twilio_machine_detection. Choose enable for an early human-or-machine verdict or detect_message_end to wait for a voicemail greeting to finish. Results arrive through the new answering_machine_detection webhook event.

Data collection value constraints: Data collection properties can now use allowed_values to name a dynamic variable containing the permitted values. The previous allowed_values_dynamic_variable field is deprecated.

Agent metadata: List agent branches responses now indicate when a draft was created and whether it predates the branch tip. Agent and Speech Engine summaries now require voice_id.

Knowledge base source URLs: Agent RAG query chunks now include required nullable source_url, containing the tracked URL for URL documents.

Agent response attachments: agent_response WebSocket events now include optional attachments. Each attachment requires url and name and can include mime_type.

Workspaces

API key platform limits: Service account API key responses can now include enterprise platform_limits for credits, Professional Voice Clones, and TTS, Dubbing, and Music concurrency.

SDK releases

JavaScript SDK

v2.67.0 - Added dynamicVariableParams to conversation listing and text search requests. Regenerated types for Twilio answering machine detection, data collection value constraints, API key platform limits, agent branch draft metadata and knowledge base source URLs.

Python SDK

v2.67.0 - Added dynamic_variable_params to synchronous and asynchronous conversation listing and text search methods. Regenerated types for Twilio answering machine detection, data collection value constraints, API key platform limits, agent branch draft metadata and knowledge base source URLs. The real-time text-to-speech client no longer passes the internal omit sentinel to the connection.

ElevenLabs CLI

v1.2.0 - Added the elevenlabs say command and its agent skill, regenerated API command bindings and made eleven_v3 the default TTS model.

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Re-exported MessageAttachment from @elevenlabs/client so consumers can import the type directly.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Forwarded agent-response file attachments to onMessage and added the optional webRtc.singlePeerConnection session setting. The widget also localizes greeting buttons and fixes dropped or duplicated text-chat responses.

API

ElevenAgents

Phone keypad input: Agent conversation configuration now supports optional dtmf_input_settings for DTMF input. Configure the digit timeout, use # as a terminator and redact keypad entries from transcripts, logs and analysis.

Conversation attachment history: Get conversation transcript entries now include file_inputs, an array containing every attached file's ID, name, MIME type and URL.

Agent tag filtering: List agents now accepts an optional tags query parameter. Repeat the parameter to match agents with any of the supplied tags.

Batch campaign context: Get conversation now identifies batch calls with nullable campaign metadata containing the required campaign_id and campaign_lead_id.

Music

Waveform visual data: Generate music detailed, Stream music with details and Upload music now accept optional with_waveform_visual (boolean, default false). Opted-in responses return a low-resolution waveform_visual integer array with four samples per second.

Dubbing

Project pagination behavior: List Dubbing projects and List language targets now clamp page_size to 1–100. Project status filtering includes queued, and both endpoints document using next_cursor to retrieve the next page.

SDK releases

JavaScript SDK

v2.66.0 - Added fileIds to sendMultimodalMessage, allowing one message to include multiple file attachments.

Python SDK

v2.66.0 - Added file_ids to send_multimodal_message, allowing one message to include multiple file attachments.

Android SDK

v0.12.2 - Preserved the onUnhandledClientToolCall callback when wiring session events. Unhandled client tool calls that expect a response now receive the automatic error response instead of remaining unresolved.

ElevenLabs CLI

v1.1.0 - Regenerated API command bindings and added command-scoped request attribution for ElevenAgents workflows. These commands append cmd/<command> to the CLI User-Agent, allowing requests from workflows such as agents push and agents create to be distinguished.

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added fileIds to sendMultimodalMessage. The client sends both file and files fields for compatibility, and React exposes the updated client API.

API

ElevenLabs CLI

The ElevenLabs CLI v1.0.0 is now available. Every ElevenLabs API operation is available as a subcommand, with JSON, table, YAML and CSV output, automatic pagination and shell completion.

The CLI also provides local configuration workflows for ElevenAgents. Store agent, tool and test configurations as files, synchronize them with push and pull commands, manage branches, run tests, install ElevenLabs UI components and select a data residency region.

Procedures

Procedures are now generally available in ElevenAgents. A procedure contains task-specific instructions and a trigger that determines when they apply. During a conversation, the agent loads the relevant procedure, allowing one agent to handle distinct tasks without placing every instruction in its system prompt.

Use free-form procedures when the agent can adapt the wording or order of instructions. Use structured procedures when steps must run in a defined order. Both types can be used alongside workflows on the same agent.

ElevenAgents

Conversation triage tickets: Added APIs for creating tickets about agent performance, creating manual follow-up tickets, listing tickets, assigning workspace members, updating status and assignee, and adding ticket-level or turn-level comments.

Conversation observability: Realtime clients can receive a context_usage event after each completed agent turn. The event reports the model, prompt token count and model context limit.

SDK Releases

JavaScript SDK

v2.65.0 - Added clients and types for conversation triage tickets and lightweight conversation summaries. Conversation APIs add procedure, invalid-tool-call and sort filters; procedure APIs add agent version selection; topic APIs add evaluation detail controls and frustration sorting. The release also adds knowledge base refresh frequency, test environment and alerting integration fields, plus realtime Dubbing message types.

Python SDK

v2.65.0 - Added clients and types for conversation triage tickets and lightweight conversation summaries. Conversation APIs add procedure, invalid-tool-call and sort filters; procedure APIs add agent version selection; topic APIs add evaluation detail controls and frustration sorting. The release also adds knowledge base refresh frequency, test environment and alerting integration fields, and preserves repeated multipart fields for Speech to Text keyterms and Dubbing webhook_ids.

Swift SDK

v3.3.0 - Added typed client error parsing, expects_response handling and type-safe tool results. WebRTC startup now becomes ready when the agent joins instead of waiting for audio-track subscription. The release also hardens URL handling, suppresses unknown-event noise and removes realtime-thread allocation from software mute processing.

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - WebRTC output capture now uses self-hosted AudioWorklet paths under strict content security policies. Worklet caching includes the requested source, preventing an earlier inline URL from replacing a self-hosted path. React Native setup errors now point to @elevenlabs/react-native.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added optional webRtc.iceTransportPolicy. Set it to relay to restrict WebRTC ICE candidates to TURN relays on networks that block direct UDP traffic.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added the onContextUsage callback and ContextUsageEvent type. React exposes the callback through useConversation. React Native now resolves through its package export condition instead of browser-global detection.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added first-message rich content with attributed button responses, onExternalAgentDisconnected and onMCPToolApprovalRequest. WebRTC now uses the selected input device, preserves mute state when switching microphones and fails setup if initiation data cannot be sent. The widget adds concurrency queue status, optional language selection on the collapsed trigger and first-message rendering for agents that support both text and voice.

API

ElevenCreative

Image, video and speech generation APIs: New asynchronous Flows APIs create, list and retrieve video, image and speech generations. Generation responses expose pending, generating, completed and failed states, and requests can include a webhook target for completion events.

Reusable media assets: New asset APIs upload, list, retrieve and delete workspace media. Image, video and audio references can use an uploaded asset, a previous generation or inline base64 data, allowing generation jobs to be chained. See References and assets.

ElevenAgents

Lightweight conversation summaries: Get conversation summary returns the generated title, transcript summary, call outcome and plain user or agent messages without loading the full transcript. The optional max_messages query parameter is an integer from 1 to 200 (default 40); longer conversations set messages_omitted to true. Useful for agents that need to preserve context.

Conversation topic pagination and sorting: Get agent conversation topics adds optional cursor pagination and sorting. Use page_size (integer, 1–100), cursor (string), sort_by (conversations, sentiment or success_rate) and sort_direction (asc or desc). Responses add has_more and nullable next_cursor.

Concurrency wait queues: Agent platform settings add optional queueing (AgentQueueingConfig). Set enabled to true to hold callers when the agent reaches its concurrency limit; wait_timeout_seconds is an integer up to 1,800 (default 180).

Turn and workflow controls: Turn configuration adds merge_with_default_ignore_terms (boolean, default false) to combine custom interruption terms with curated defaults. Soft timeouts add disable_until_first_user_message (boolean, default false), and workflow tool locators add optional schema_overrides for per-node parameter overrides.

Background audio defaults: BackgroundSoundConfig.volume now defaults to 0.15 instead of 0.6, and crossfade_loop now defaults to true instead of false. Set these fields explicitly to retain the previous behavior.

SDK Releases

JavaScript SDK

v2.64.0 - Added client.assets for uploading, listing, retrieving and deleting assets, plus client.flows for asynchronous video, image and speech generation. Turn configuration adds mergeWithDefaultIgnoreTerms. Removed CustomToolApiSchemaConfig and scimExternalId from WorkspaceGroupResponseModel.

Python SDK

v2.64.0 - Added assets and Flows clients for asynchronous video, image and speech generation, plus merge_with_default_ignore_terms in turn configuration. On-prem session setup adds extra_setup_config for passing additional setup-message fields.

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added experimental self-hosted orchestrator sessions and rich-content events. Scribe audio worklets can now be self-hosted for strict content security policies. Disconnect handling now reaches disconnected consistently and prevents late callbacks from clearing a newer session. The widget renders quick replies and links inline, restores Markdown in voice transcripts and fixes language menu positioning inside container-query layouts.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added realtime Speech to Text options for secondary languages, entity detection and background audio filtering. Added final transcript, detected entity and invalid request events; aligned timestamp word types with the live schema; and fixed microphone permission failures preventing useScribe reconnection.

API

Dubbing v2 API

Dubbing v2 is now available through the API. It translates audio and video into more than 90 languages while preserving each speaker's voice, tone and pacing.

The new project-based API keeps source transcripts and translations as editable JSON. Create a project from a file or URL, add one or more target languages, edit individual transcript segments or translations, then regenerate only the regions that changed. Follow the Dubbing quickstart to create your first dub.

ElevenAgents

Conversation guardrail filters: List conversations adds optional guardrail_types (GuardrailType[]) and custom_guardrail_names (string array) query parameters. The existing conversation hierarchy can now be filtered with optional parent_conversation_id (string).

Branch divergence status: List agent branches adds optional include_commit_status (boolean, default false). When enabled, branch summaries include nullable commits_ahead, commits_behind and merged_into_branch_id fields.

Conversation and tool controls: Client overrides now support max_duration_seconds and pronunciation_dictionary_locators. MCP server configuration adds request_meta for passing MCP _meta values to tool calls, and webhook tool timeouts now support up to 300 seconds.

Post-call analysis usage: Conversation charging adds an analysis breakdown with per-feature running totals and run snapshots. Transcript messages add triggered_guardrails, and conversation turns add optional producing_llm.

Voices

Voice search filters: Get all voices v2 adds optional gender, age, language, accent, use_cases, min_notice_period_days, include_custom_rates, include_live_moderated and high_quality query parameters.

SDK Releases

JavaScript SDK

v2.63.0 - Added realtime Speech to Text options for secondary languages, language and entity detection, background audio filtering, logging and single-use token authentication. Added final transcript, entity and invalid request events; fixed unaccepted terms dispatch, duplicate audio_format parameters and inclusive VAD bounds.

v2.62.0 - Added Dubbing bulk transcript update methods and DubbingRegenerateResponse; added conversation guardrail filters, agent branch commit status, agent alerting and MCP requestMeta support. Language target creation no longer accepts modelId.

v2.61.0 - Added conversation hierarchy filtering, post-call analysis charging, triggered guardrail metadata, branch merge conflict details, DTMF input redaction and Dubbing transcript uploads. Removed exported OpenAI realtime session types and run_subagent_* tool-result variants.

Python SDK

v2.63.0 - Added realtime Speech to Text options for secondary languages, language and entity detection, background audio filtering, logging and single-use token authentication. Added final transcript, entity and invalid request events; fixed unaccepted terms dispatch and now rejects connections without an API key or token.

v2.62.0 - Added Dubbing update_segments methods and DubbingRegenerateResponse; added guardrail_types, custom_guardrail_names, include_commit_status and MCP request_meta support. Dubbing segment updates now use request body models, and language target creation no longer accepts model_id.

v2.61.0 - Added parent_conversation_id, analysis charging, triggered guardrail metadata, branch merge conflicts, DTMF input redaction and Dubbing transcript uploads. Removed OpenAI realtime session types and run_subagent_* result models.

Android SDK

v0.12.1 - Fixed intentional text-only session disconnects being reported as ConversationStatus.ERROR.

v0.12.0 - Added ConversationConfig.useMediaStream for routing conversation audio through the Android media stream. Remote normal WebSocket closures now report DisconnectionDetails.Agent, while locally initiated closures report DisconnectionDetails.User.

MCP Server

v0.12.2 - Fixed an issue with path traversal.

API

ElevenAgents

Test-specific tool mocks: Agent simulations and unit tests can now define tool_mock_overrides, keyed by tool ID, to replace shared mocks for one test. ToolResponseMockConfig also adds optional is_error (boolean, default false) so a mock can exercise tool-failure paths.

Translated workflow messages: Literal Say node messages add text_translations, a map of language keys to TranslatedString objects. The field is optional on input and required on output, so strict response validators must accept it.

Conversation version filters: List conversations and Text search conversation messages add the optional version_id query parameter (string) to filter results by agent version.

Knowledge base system tool: Agent configuration adds a knowledge_base system tool with configurable enabled_strategies (SearchStrategy[]). Agent knowledge base RAG queries add optional use_agent_defaults (boolean, default true); set it to false to use neutral retrieval settings while retaining the agent's embedding model.

Agent voice activity detection: Agent create, update, response and workflow-override schemas add vad configuration. VADConfig currently supports background_voice_detection (boolean, default false).

WebSocket custom LLMs: CustomLLMAPIType adds websocket.

File memory controls: Widget file-input configuration adds max_files_in_memory (integer, 1–30, default 10). Files beyond the limit are summarized and released from memory; max_files_per_conversation must be -1 or at least the memory limit.

DTMF input redaction: DTMFInputConfig adds optional redact_input (boolean, default false) to replace caller keypad digits in transcripts, conversation logs and analysis. Digits repeated by the agent or passed to tools remain unchanged.

Speech to Text

Realtime entity detection: Scribe v2 Realtime adds the entity_detection query option. It accepts all, one entity type or category, or an array of types and categories, and emits detected entities in committed_transcript_entities events. See the entity detection guide for supported categories.

Voices

Voice replication across data residencies: Added POST /v1/voices/{voice_id}/replicate-to-isolated-environment for copying an Instant Voice Clone or Voice Design voice to a workspace in the same consolidated billing group. Requests require target_workspace_id (string) and accept preserve_voice_id (boolean, default true); responses return the target voice_id.

Character metadata: Character responses add recommended_voice_ids (string array, default []), metadata adds nullable role (CharacterRole), and CharacterAge replaces middle-aged with middle_aged.

SDK Releases

Packages

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added onAgentReasoningResponsePart, which receives streaming agent reasoning events with start, delta and stop types.

@elevenlabs/[email protected] and @elevenlabs/[email protected] - Added show_resize_button and the show-resize-button embed attribute. Set either to false to hide the widget header's expand and collapse control.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Added the enableLogging option to Scribe.connect and useScribe. Setting it to false starts an enterprise zero-retention session. The Scribe session_started type now reports enable_logging; the unused disable_logging field was removed.

MCP Server

v0.12.1 - Includes the v0.11 tool updates and a dependency refresh that resolves 44 security alerts.

v0.11.0 - Added simulate_conversation with configurable simulated-user behavior, evaluation criteria, transcript output and pass/fail analysis. The speech_to_text tool now defaults to scribe_v2, and voice_clone now sends audio file contents instead of path strings.

API

ElevenAgents

WhatsApp branch and environment overrides: Outbound WhatsApp messages and outbound calls now honor branch_id and environment passed in conversation_initiation_client_data. Overrides are validated before anything is sent — an unknown branch or environment is rejected with an error — and persist for the conversation, so a customer's reply resumes on the requested branch and environment. See the WhatsApp guide for examples.

Procedure APIs: Added eight public, branch-scoped operations for managing procedures. You can create, list, retrieve and remove procedures; create, retrieve or discard draft changes; and compile procedure drafts into an agent workflow. Procedure creation accepts name, content, type and trigger. An empty trigger creates a sub-procedure, while an omitted or null trigger is derived from the procedure content.

Knowledge base crawl jobs: Added APIs to create, list, inspect and cancel crawl jobs. Create requests require url (string) and support crawl depth, page limits, URL patterns, sitemap URLs, folder placement, automatic synchronization and removal of unavailable documents.

Bulk knowledge base management: Added bulk dependency checks and bulk deletion for 1–20 unique document or folder IDs. Dependency checks support pagination and dependent_type; deletion returns an independent success or failure for each ID. Setting force: true also removes dependencies and recursively deletes non-empty folders.

Batch call result export: Added Export batch call results to download recipients and conversation results from a terminal batch as CSV.

Conversation filters and analysis data: List conversations and Text search conversation messages add optional visited_agent_ids and visited_agent_branch_ids array filters, each limited to 50 values. List requests also accept data_collection_ids and evaluation_criteria_ids to include corresponding results in each conversation summary. Summaries add optional data_collection_results, evaluation_criteria_results and tag_ids.

Referenced analysis configuration: Agent platform settings add the optional, nullable analysis_items field, typed as AgentAnalysisItems, to attach evaluation and data-collection items by reference. A null value indicates that the agent still uses the legacy inline fields; an empty value indicates a migrated agent with no attached items.

Procedure dependency metadata: Procedure references add trigger (string) and arrays for referenced_tool_ids, referenced_kb_ids, referenced_procedure_ids and referenced_dynamic_variables.

WebRTC token response: Get WebRTC token responses now require conversation_id (string) in addition to token. This is a breaking schema change for clients that validate or deserialize the response shape.

Knowledge base RAG results: KnowledgeBaseRagToolResultModel adds chunks (KnowledgeBaseRagChunkModel[]) for RAG-result-in-tool-result mode.

Studio

Draft project status: ProjectCreationMetaResponseModel.status adds draft. The project creation type now uses the shared ProjectCreationMetaType enum with the same existing values.

Voices

Voice accents: Added GET /v1/voices/accents to list the accents available in the shared voice library. Results can be filtered by language and model_id, and each accent returns its accent, language, code and human-readable name. The accent value can be passed to the accent query parameter on Get shared voices to filter shared voices.

Workspaces

Webhook usage reporting: WebhookUsageType adds Flows.

SDK Releases

JavaScript SDK

v2.60.0 - Added voices.accents.get to list voice accents in the shared voice library, along with the ElevenAgents procedure, knowledge base crawl job and bulk management APIs and the referenced analysis configuration from the latest schema.

Python SDK

v2.60.0 - Added the voice accents, ElevenAgents procedure, knowledge base crawl job and bulk management APIs from the latest schema. Also fixed AsyncConversation startup so it no longer blocks, and corrected list-of-primitive multipart fields to be sent as repeated form fields.

Packages

@elevenlabs/[email protected] - Fixed onAudioAlignment on WebRTC transports by forwarding alignment metadata from audio messages while leaving audio playback on the LiveKit track.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Updated dependencies to include the WebRTC alignment fix from @elevenlabs/[email protected].

API

Music

Music Finetunes API: Added API support for managing Music Finetunes. Five endpoints let you create a Finetune from uploaded audio, list accessible Finetunes, retrieve training status, update metadata and visibility, or delete a Finetune. All music generation SDK methods now expose finetune_id.

ElevenAgents

Resolve external conversation references: Added GET /v1/convai/conversations/resolve to resolve a Slack message URL or Zendesk ticket URL to a conversation. Requests require agent_id and reference query parameters (strings).

Query agent knowledge bases: Added POST /v1/convai/agents/{agent_id}/knowledge-base/rag-query for read-only RAG queries against an agent's knowledge base. The request requires query (string), accepts optional branch_id (string), and returns ranked chunks with document, chunk, text and vector-distance metadata. This endpoint is excluded from generated SDKs.

Knowledge base content formats: Knowledge base file, text and URL responses add optional content_format (html | markdown, default html).

Agent voice configuration: TTS override schemas add optional client-overridable model_id (TTSConversationalModel). The default for enable_phoneme_tags changes from false to true.

Faster backup LLM cascade: The default cascade_timeout_seconds changes from 8 to 4 seconds. The allowed range remains 2–15 seconds.

Tool parameter value sources: Object tool parameters can now use dynamic_variable, constant_value or is_omitted, matching existing scalar and array controls. Constant schema overrides now accept arbitrary arrays and objects.

Full-response guardrails: Custom guardrails add optional evaluate_full_response_only (boolean, default false) to evaluate the complete non-TTS response once. This option requires blocking mode.

Twilio call recording: Telephony call configuration adds optional twilio_call_recording_enabled (boolean, default false). Recordings are stored in the connected Twilio account and the setting is ignored by other providers.

Widget file-upload text: Widget text configuration adds optional labels and error messages for attaching and removing files, upload failures, unsupported types, oversized files, file-count limits and the typing indicator.

Knowledge base sync status: Knowledge base file responses add optional auto_sync_info and refresh_status (FileRefreshStatus). Folder responses add required document_count (integer), a recursive count of non-folder documents capped at 1000.

Conversation text search filters: Text search conversation messages adds optional exclude_statuses and termination_reasons array query parameters.

LLM options: Agent LLM configuration adds gpt-5.6-sol, gpt-5.6-terra and gpt-5.6-luna. LLMReasoningEffort adds max.

WhatsApp typing indicators: WhatsApp account request and response schemas add optional enable_typing_indicator (boolean).

Speech to Text

Single-use batch authentication: Convert speech to text adds optional token (string) as an alternative to API key or bearer authentication. Create the single-use token with POST /v1/single-use-token/batch_scribe; it expires after 15 minutes.

Realtime model and language options: Realtime Speech to Text adds scribe_v2_realtime_turbo and scribe_v2_realtime_lite model IDs, plus secondary_languages for additional transcription languages.

Realtime client reliability: @elevenlabs/client now reports microphone setup and malformed server errors through RealtimeEvents.ERROR, releases microphone resources after setup failures and connection closure, and drops late microphone frames after socket teardown.

Text to Speech

Archived pronunciation dictionaries: List pronunciation dictionaries adds optional include_archived (boolean, default true).

Workspaces

Webhook event subscriptions: Update workspace webhook requests add optional events (WorkspaceWebhookEventType[]) with voice_library_removal_notice, speech_to_text and agent_qa. Webhook responses can return the subscribed event list.

Usage analytics dimensions: Get workspace usage adds surface and actor to the group_by values.

Clear service account limits: Update service account API key accepts character_limit: "clear" to remove a monthly character limit, alongside "no_update" and integer values.

Alerting webhook references: Alerting webhook notifier schemas now identify a workspace webhook with required webhook_id (string) instead of embedding url, method and headers.

SDK Releases

JavaScript SDK

v2.59.0 - Added music.finetunes methods to list, create, get, update and delete Music Finetunes. Also added conversation reference resolution, archived pronunciation dictionary filtering and single-use Speech to Text tokens.

Python SDK

v2.59.0 - Added Music Finetunes and conversation reference resolution from the latest API schema. OnPremInitiationData also adds typed post_call_transcription_webhook and post_call_audio_webhook configurations with optional hmac_secret for signed post-call webhooks.

Packages

@elevenlabs/[email protected] - Improved Scribe Realtime error delivery and microphone cleanup during setup failures, client or server closure, and late audio frames.

@elevenlabs/[email protected], @elevenlabs/[email protected], @elevenlabs/[email protected] and @elevenlabs/[email protected] - Updated dependencies to include the @elevenlabs/[email protected] Scribe Realtime fixes.

Android SDK

v0.11.1 - Prevented R8 and ProGuard from obfuscating outgoing WebSocket and WebRTC event fields such as text and type.

CLI

@elevenlabs/[email protected] - Updated default agent templates to use scribe_realtime instead of the removed elevenlabs ASR provider. Minimal templates continue to use the server default.

API