Create a new agent
Create a new agent. By default, creates a new version (v1) and a draft. If is_draft is true, creates only a draft.
Autorizaciones
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Parámetros de consulta
If true, the agent is created as a draft.
Cuerpo
The voice id to be used by the agent
The language used by the agent. Use multi for the agent using both english and spanish in the same call (multi will expand to more languages in the future)
es, en, multi, ar, be, bn, bs, bg, ca, hr, cs, da, nl, et, fi, nl-BE, fr, de, de-CH, el, he, hi, hu, id, it, ja, kn, ko, lv, lt, mk, ms, mr, no, fa, pl, pt, ro, ru, sr, sk, sl, sv, tl, ta, te, tr, uk, ur, vi The name of the agent
The description given to the agent
The prompt given to the agent
The greeting message said by the agent to the caller at the beginning of the call
The time in seconds before the agent starts to speak
Controls the output volume of the agent's voice. 50 is the default (no change), 0 is silent, 100 is maximum gain.
0 <= x <= 100The maximum number of reminders that can be sent to the caller. If not provided, the default value is 3. When reached the last reminder, the call gets hang up. Set this value to 0 to disable reminders
x >= 0The frequency in milliseconds to send reminders to the caller. If not provided, the default value is 7500. Set this value to 0 to disable reminders
x >= 1000The list of messages to send to the caller as reminders. If not provided, the default value is an empty list which makes the model to generate the reminder messages by itself. Please, note that when using an empty list, the reminder_frequency_ms can be a little bit higher than specified since it has to generate the messages
The list of knowledge base ids available for the agent to use
The list of knowledge item ids directly available for the agent to use
Conversation mode. 'free' (default) is a single-prompt agent. 'path' is a structured graph; setting mode='path' REQUIRES conversational_path_id to point at an already-created path (call diga_create_path first to obtain it).
free, path ID of a conversational path. Required when mode='path'; ignored when mode='free'. Create the path first via diga_create_path, then pass its id here.
The name of the created version
The list of integrations and their tools to assign to the agent
Built-in call-tools to attach to the agent. Each list element is FLAT (no nested per-variant object): the discriminator is the top-level type field, with the variant's other fields as siblings. Example: [{"type": "transfer_call", "name": "transfer_to_owner", "target_number": "+34600112233", "description": "Transfer when the user asks to talk to the owner"}, {"type": "hang_up", "name": "end_call"}]
Transfer a call to another number.
Wire shape (FLAT, all fields top-level):
- TransferCallToolSchema
- HangUpToolSchema
Agent-level dynamic variables for template interpolation
The background audio to be used in the calls that this agent handles
call_center, cafe, road_traffic, park, office, car_garage List of flow IDs to assign to the agent
Detect when an outbound call reaches a voicemail
What the agent does when a voicemail is detected
leave_message, hang_up Message the agent leaves on the voicemail. Required when voicemail_detection is on with action 'leave_message'. Supports {{dynamic_variables}}.
Detect and handle automated phone menus (IVR) on outbound calls
What the agent does when a phone menu is detected
navigate, hang_up How the agent navigates the menu: towards a plain-language goal or pressing a fixed key sequence
goal, key_sequence Goal the agent navigates the menu towards, e.g. 'Reach a human in billing'. Required when navigating in 'goal' mode. Supports {{dynamic_variables}}.
DTMF keys pressed in order (0-9, *, #), e.g. '1,3,5'. Required when navigating in 'key_sequence' mode.
Respuesta
Successful Response
The unique identifier for the agent
The date and time when the agent was last updated
The date and time when the agent was created
The version id of the agent
The version number of the agent
The date and time when the version was created
The voice object used by the agent
The language used by the agent. Use multi for the agent using both english and spanish in the same call (multi will expand to more languages in the future)
es, en, multi, ar, be, bn, bs, bg, ca, hr, cs, da, nl, et, fi, nl-BE, fr, de, de-CH, el, he, hi, hu, id, it, ja, kn, ko, lv, lt, mk, ms, mr, no, fa, pl, pt, ro, ru, sr, sk, sl, sv, tl, ta, te, tr, uk, ur, vi The mode of the agent. Can be 'free' or 'path'
free, path The version name of the agent
Indicates if the agent is a draft version
Indicates if the agent is published
The name of the agent
The description given to the agent
The prompt given to the agent
The greeting message said by the agent to the caller at the beginning of the call. If set, the agent will always speak first both in inbound and outbound calls
The time in seconds before the agent starts to speak
Controls the output volume of the agent's voice. 50 is the default (no change), 0 is silent, 100 is maximum gain.
0 <= x <= 100The maximum number of reminders that can be sent to the caller. If not provided, the default value is 3. Set this value to 0 to disable reminders
x >= 0The frequency in milliseconds to send reminders to the caller. If not provided, the default value is 7500.
The list of messages to send to the caller as reminders. If not provided, the default value is an empty list which makes the model to generate the reminder messages by itself. Please, note that when using an empty list, the reminder_frequency_ms can be a little bit higher than specified since it has to generate the messages
The list of knowledge bases available for the agent to use
The list of knowledge items directly available for the agent to use
The id of the conversational path used by the agent.
The list of integrations and their assigned tools for the agent
The list of call-tools (built-in tools) assigned to the agent
Transfer a call to another number.
Wire shape (FLAT, all fields top-level):
- TransferCallToolSchema
- HangUpToolSchema
Agent-level dynamic variables for template interpolation. Overridden by call-level variables.
The background audio to be used in the calls that this agent handles
call_center, cafe, road_traffic, park, office, car_garage Detailed flow info assigned to the agent
The ID of the agent template this agent was created from
Detect when an outbound call reaches a voicemail
What the agent does when a voicemail is detected
leave_message, hang_up Message the agent leaves on the voicemail
Detect and handle automated phone menus (IVR) on outbound calls
What the agent does when a phone menu is detected
navigate, hang_up How the agent navigates the menu: towards a goal or pressing a fixed key sequence
goal, key_sequence Plain-language goal the agent navigates the menu towards
Comma-separated DTMF keys pressed in order, e.g. '1,3,5'
List of issues that would prevent launching/publishing. Only present for draft versions on dashboard endpoints.

