Skip to main content
PATCH
Update Agent

Authorizations

string
required
Your Agni API key used to authenticate the request. Pass it in the X-Api-Key header. Find or rotate it from your Ravan AI account settings.

Request

Example curl:

Path Parameters

string
required
The unique ID of the resource in the path. Use the ID returned by the related create or list endpoint.

Body

string
The public-facing display name for the agent. This is the name users may see or hear during conversations.
string
The LLM model identifier used by the agent for reasoning and responses.
string
The speech-to-speech model identifier used for realtime voice conversations.
string
The voice ID used for the agent speech output.
number
Controls response randomness for the agent model. Lower values are more deterministic; higher values are more varied.
integer
The amount of silence, in milliseconds, before the agent sends a reminder or prompt.
integer
The maximum number of reminder messages the agent can send during a call.
string
The background sound profile to play during calls. Leave empty to disable ambient audio.
number
The volume level for the selected ambient sound. Use a decimal value from 0 to 1, where 0 is muted and 1 is full volume.
boolean
Whether the call should end automatically after a long silence from the user.
integer
The maximum allowed call duration in milliseconds. The call ends when this limit is reached.
integer
The maximum time, in milliseconds, to let an outbound call ring before it is treated as unanswered.
boolean
Whether the agent should detect voicemail during outbound calls.
string
The message the agent should leave when voicemail is detected.
integer
The maximum time, in milliseconds, to wait for voicemail detection before continuing the call flow.
boolean
Whether to transfer the call to a backup number on failure.
string
The backup number to transfer the call to.
object
Interactive voice response settings for the agent, including keypad options and routing behavior.
string
The model used to generate post-call analysis, such as summaries, sentiment, and disposition.
string
The prompt used to generate the post-call summary. Include the outcome, details, and next steps you want the AI to extract.
object
The tools enabled for the agent. Include tool IDs and any configuration needed for the agent to call them.
object
Integration settings for the agent, such as connected scheduling, CRM, or webhook configuration.
string
The knowledge base reference or content available to the agent during conversations.
string
The first message the agent says when the conversation starts. Keep it short and natural for voice calls.
string
Who speaks first when the call starts. Use the value supported by the agent configuration, such as agent or user.

Response

boolean
Whether the request succeeded.
string
Human-readable status message.
object
Returned data.

Authorizations

X-Api-Key
string
header
required

Path Parameters

id
string
required

Agent ID

Body

application/json
widget_settings_id
string
name
string
agent_name
string
model
string
s2s_model
string
voice_id
string
temperature
number<double>
reminder_trigger_ms
integer<int32>
reminder_max_count
integer<int32>
ambient_sound
string
ambient_sound_volume
number<double>
end_call_after_silence
boolean
max_call_duration_ms
integer<int32>
ring_duration_ms
integer<int32>
enable_voicemail_detection
boolean
voicemail_message
string
voicemail_detection_timeout_ms
integer<int32>
enable_emergency_fallback
boolean
emergency_fallback_number
string
ivr_options
object
post_call_analysis_model
string
analysis_summary_prompt
string
selected_tools
object
integrations
object
knowledge_base
string
begin_message
string
start_speaker
string

Response

200 - application/json

Successful response

success
boolean
message
string
data
object