Deletes a note from a conversation.

Deletes multiple users and all linked conversations and messages.

Sends a direct chat message. Details at

Returns an array with all the articles or a single article.

Returns all the conversations. Each conversation includes the last message of the conversation.

Returns the last message in a conversation.

Returns the conversations created after the given date/ID or with a message created after the given date/ID. Each conversation includes the last message of the conversation.

Returns the user conversations created after the given date/ID or with a message created after the given date/ID. Each conversation includes the last message of the conversation.

Returns a setting saved in the Settings area.

Returns the conversations of a user. Each conversation includes the last message of the conversation.

Translates multiple strings via Google Translate.

Deletes an existing message.

Updates an existing message.

Creates a new conversation.

Sends a Push notification to an agent, a group of agents, or all agents. Push notifications must be enabled in the settings area.

Returns the conversations matching the search.

Sends a generic email to an email address. The sender email and name are the ones set in Settings > Notifications > Email settings.

Sends a text message to a user or agent. If the template argument is true, the message is translated automatically.

Assigns or update the agent assigned to a conversation.

Updates the status code of a conversation.

Sends a message to a WhatsApp number.

Starts consent for a private avatar group and returns a URL to open in a browser (v3 POST /avatars/{group_id}/consent).

Generate a video featuring one or multiple scenes with avatars, talking photos, and various audio and background options. This action combines multiple scenes into a cohesive video.

Creates a single scene configuration for an avatar video.

Generates a WebM video with transparent background and an avatar speaking the input text.

Returns up to 3 voices matching a natural language description (v3 POST /v3/voices). Use seed for different batches.

Synthesize speech from text with a starfish-compatible voice (v3). Use List Voices with engine=starfish to pick a voice.

Returns one avatar group: name, previews, looks count, training status (v3 GET /avatars/{group_id}).

Reads session status, video_id, and recent messages (v3 GET /video-agents/{session_id}).

Paginated looks (avatar_id for video APIs); filter by group, type, or ownership (v3 GET /avatars/looks).

Returns a paginated list of voices (v3). Filter by type, engine, language, or gender. Use engine=starfish for voices compatible with Generate Speech.

Get the status and details of a video translation job (v3).

Sends a follow-up or revision message to a chat-mode session (v3 POST /video-agents/{session_id}).

Renames a look (photo avatar and digital twin only per API; v3 PATCH /avatars/looks/{look_id}).

Deletes a user and all linked conversations and messages.

Sends a message to Dialogflow and add the Dialogflow response to an existing conversation as a new message.

Returns an array with all the agents with at least one message in the conversation.

Returns a conversation and the messages of the conversation.

Returns an external setting. External settings are saved on a dedicated database row.

Returns a subsetting of another setting.

Returns the messages of a conversation created after the given date/ID.

Returns the internal notes of a conversation.

Returns the last agent who answered a conversation.

Logins a user or agent. A user can log in in two ways: via email and password, or via user ID and token. The last value of the response is the encrypted login data ready to be stored in the sb-login cookie. Use the function SBF.loginCookie(response[1]); to store it. Returns false if login is unsuccessful.

Adds a new message to an existing conversation.

Sends a message to Messenger or Instagram.

Sends a message to OpenAI and return the response, optionally add the response to an existing conversation as a new message.

Returns the articles matching the search.

Returns the conversations of the given user ID that matches the search terms.

Sends an email to an existing user using the email templates of the Settings > Notifications area.

Sends a message to Telegram.

Updates the department of a conversation.

Updates the details of an existing user.

Triggers when a new video event happens. You can customize which event(s) will trigger.

Starts asynchronous avatar creation from a prompt or media URLs (v3 POST /avatars). Photo/digital twin flows may require image, video, or consent_video—check HeyGen docs for your type.

Starts a proofread session: editable subtitles are extracted before final render (v3 POST /video-translations/proofreads).

Starts a HeyGen Video Agent session from a text prompt (v3). Poll session status or use webhooks for the final video.

Creates a video from an avatar or image (v3 POST /videos). Avatar III still uses legacy v1/v2; use v2 multi-scene for complex timelines.

Permanently deletes a video and its files (v3 DELETE /videos/{video_id}). Cannot be undone.

Generate customized videos from templates. Templates offer a rich, versatile content format for your workflow, allowing you to customize text, images, videos, audios, voices, and avatars.

Starts final translated video generation from an approved proofread session (v3 POST .../proofreads/{id}/generate).

Returns details for one look (avatar_id for videos): engines, orientation, previews, training status (v3 GET /avatars/looks/{look_id}).

Paginated avatar groups (characters); each group has one or more looks (v3 GET /avatars).

Paginated list of videos in the account (v3 GET /videos). Filter by folder or title substring.

Get a public shareable link for a HeyGen video that can be easily distributed and viewed by others.

Get the status and details of a specific video (v3).

Creates a video translation job: target languages with voice cloning and lip-sync (v3 POST /video-translations).

Upload an image, video, or audio file to HeyGen and receive a storage key for use in other operations.