September 21, 2026 | ElevenLabs Documentation

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.

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.

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.

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.

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.

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.

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.

New endpoints

List phone numbers page - GET /v1/convai/v2/phone-numbers
- Added optional page_size (integer, 1–1,000, default 100) and cursor (nullable string) pagination parameters.
- Added optional search, label, phone_number, provider, supports_outbound, agent_id, branch_id, sort_by and sort_direction filters. sort_by accepts label or phone_number; sort_direction defaults to asc.
- The response contains required phone_numbers (array), optional nullable next_cursor (string) and has_more (boolean, default false).

Create agent, Get agent and Update agent
- Prompt configuration adds enable_parallel_tool_calls (boolean, default true). Workflow overrides accept a nullable boolean.
- The LLM enum adds gpt-6-astra, including conversation initiation overrides.
- Alerting integration notifiers now use integration_type to select pagerduty or slack. Slack notifiers require connection_id and channel_id (strings); channel_id accepts 1–64 characters.
List conversations and Resolve conversation reference
- data_collection_params no longer accepts the missing operator. Use an empty value to match conversations where the field was not collected.
- An empty value in dynamic_variable_params matches conversations where the variable was not set.
Create Speech Engine, Get Speech Engine and Update Speech Engine
- Added cascade_timeout_seconds (number, 2–15). Create requests default to 4, update requests accept a nullable value and responses require the field.

Create image generation - POST /v1/flows/image
- Added request variants for gpt-image-2.5-flare and gpt-image-2.5-sunburst.
- Both variants require prompt and model_id, accept up to 10 images, and add resolution (1K, 2K or 4K, default 1K).
- quality accepts low, medium, high, xhigh or max and defaults to high. aspect_ratio accepts auto and 14 fixed ratios and defaults to 16:9.

Create dubbing project - POST /v1/dubbing/project
- Multipart keyterms accepts repeated fields or a single JSON array.
- Multipart webhook_ids accepts repeated fields, a single JSON array or a comma-separated string.
Get workspace resource, Share workspace resource and Unshare workspace resource
- WorkspaceResourceType adds dubbing_project.

Create crawl job - POST /v1/convai/knowledge-base/crawl
- Deprecated max_depth (integer, default 3). The field is now a no-op and no longer has minimum or maximum constraints.
- Crawl job metadata no longer requires max_depth.

Multi-context WebSocket audio and final messages now define is_final_audio_for_turn as occurring after all buffered bytes for the turn have been delivered.