Update Agent

Patches an Agent settings

Headers

xi-api-key

string

Your API key. This is required by most endpoints to access our API programatically. You can view your xi-api-key using the 'Profile' tab on the website.

Path Parameters

agent_id

string

required

The id of an agent. This is returned on agent creation.

Body

application/json

conversation_config

object

Conversation configuration for an agent

platform_settings

object

Platform settings for the agent are all settings that aren't related to the conversation orchestration and content.

secrets

object[]

A list of secrets for the agent. Can be used to add new secrets or update and delete the existing ones

• ConvAINewSecretConfig
• ConvAIStoredSecretConfig

Show child attributes

secrets.type

enum<string>

required

Available options:

new

secrets.name

string

required

secrets.value

string

required

name

string

A name to make the agent easier to find

Response

200 - application/json

agent_id

string

required

name

string

required

conversation_config

object

required

Show child attributes

conversation_config.agent

object

Show child attributes

conversation_config.agent.prompt

object

Show child attributes

conversation_config.agent.prompt.prompt

string

default:

conversation_config.agent.prompt.llm

enum<string>

default: gemini-1.5-flash

An enumeration.

Available options:

gpt-4o-mini,

gpt-4o,

gpt-4,

gpt-4-turbo,

gpt-3.5-turbo,

gemini-1.5-pro,

gemini-1.5-flash,

gemini-1.0-pro,

claude-3-5-sonnet,

claude-3-haiku,

grok-beta,

custom-llm

conversation_config.agent.prompt.temperature

number

default: 0

conversation_config.agent.prompt.max_tokens

integer

default: -1

conversation_config.agent.prompt.tools

object[]

• WebhookToolConfig
• ClientToolConfig

Show child attributes

conversation_config.agent.prompt.tools.type

enum<string>

required

Available options:

webhook

conversation_config.agent.prompt.tools.name

string

required

conversation_config.agent.prompt.tools.description

string

required

conversation_config.agent.prompt.tools.api_schema

object

required

Configuration for a webhook that will be called by an LLM tool.

Show child attributes

conversation_config.agent.prompt.knowledge_base

object[]

Show child attributes

conversation_config.agent.prompt.knowledge_base.type

enum<string>

required

Available options:

file,

url

conversation_config.agent.prompt.knowledge_base.name

string

required

conversation_config.agent.prompt.knowledge_base.id

string

required

conversation_config.agent.prompt.custom_llm

object

Show child attributes

conversation_config.agent.prompt.custom_llm.url

string

required

conversation_config.agent.prompt.custom_llm.model_id

string

conversation_config.agent.prompt.custom_llm.api_key

object

Used to reference a secret from the agent's secret store.

Show child attributes

conversation_config.agent.first_message

string

default:

conversation_config.agent.language

string

default: en

conversation_config.asr

object

Show child attributes

conversation_config.asr.quality

enum<string>

default: high

An enumeration.

Available options:

high

conversation_config.asr.provider

enum<string>

default: elevenlabs

An enumeration.

Available options:

elevenlabs

conversation_config.asr.user_input_audio_format

enum<string>

default: pcm_16000

An enumeration.

Available options:

pcm_16000,

pcm_22050,

pcm_24000,

pcm_44100,

ulaw_8000

conversation_config.asr.keywords

string[]

conversation_config.turn

object

Show child attributes

conversation_config.turn.turn_timeout

number

default: 7

conversation_config.turn.mode

enum<string>

default: turn

An enumeration.

Available options:

silence,

turn

conversation_config.tts

object

Show child attributes

conversation_config.tts.model_id

enum<string>

default: eleven_turbo_v2

An enumeration.

Available options:

eleven_turbo_v2,

eleven_turbo_v2_5

conversation_config.tts.voice_id

string

default: cjVigY5qzO86Huf0OWal

conversation_config.tts.agent_output_audio_format

enum<string>

default: pcm_16000

An enumeration.

Available options:

pcm_16000,

pcm_22050,

pcm_24000,

pcm_44100,

ulaw_8000

conversation_config.tts.optimize_streaming_latency

enum<integer>

default: 3

An enumeration.

Available options:

0,

1,

2,

3,

4

conversation_config.tts.stability

number

default: 0.5

conversation_config.tts.similarity_boost

number

default: 0.8

conversation_config.tts.pronunciation_dictionary_locators

object[]

Show child attributes

conversation_config.tts.pronunciation_dictionary_locators.pronunciation_dictionary_id

string

required

conversation_config.tts.pronunciation_dictionary_locators.version_id

string

required

conversation_config.conversation

object

Show child attributes

conversation_config.conversation.max_duration_seconds

integer

default: 600

conversation_config.conversation.client_events

enum<string>[]

Available options:

conversation_initiation_metadata,

asr_initiation_metadata,

ping,

audio,

interruption,

user_transcript,

agent_response,

agent_response_correction,

client_tool_call,

internal_vad_score,

internal_turn_probability,

internal_tentative_agent_response

metadata

object

required

Show child attributes

metadata.created_at_unix_secs

integer

required

secrets

object[]

required

Show child attributes

secrets.type

enum<string>

required

Available options:

stored

secrets.secret_id

string

required

secrets.name

string

required

platform_settings

object

Show child attributes

platform_settings.auth

object

Show child attributes

platform_settings.auth.enable_auth

boolean

default: false

platform_settings.auth.allowlist

object[]

Show child attributes

platform_settings.auth.allowlist.hostname

string

required

platform_settings.auth.shareable_token

string

platform_settings.evaluation

object

Settings to evaluate an agent's performance. Agents are evaluated against a set of criteria, with success being defined as meeting some combination of those criteria.

Show child attributes

platform_settings.evaluation.criteria

object[]

Show child attributes

platform_settings.evaluation.criteria.id

string

required

platform_settings.evaluation.criteria.name

string

platform_settings.evaluation.criteria.type

enum<string>

default: prompt

Available options:

prompt

platform_settings.evaluation.criteria.conversation_goal_prompt

string

required

Maximum length: 800

platform_settings.widget

object

Show child attributes

platform_settings.widget.variant

enum<string>

default: full

An enumeration.

Available options:

compact,

full,

expandable

platform_settings.widget.avatar

object

• Avatar
• Avatar
• Avatar

Show child attributes

platform_settings.widget.avatar.type

enum<string>

default: orb

Available options:

orb

platform_settings.widget.avatar.color_1

string

default: #2792dc

platform_settings.widget.avatar.color_2

string

default: #9ce6e6

platform_settings.widget.feedback_mode

enum<string>

default: none

An enumeration.

Available options:

none,

during,

end

platform_settings.widget.custom_avatar_path

string

platform_settings.widget.bg_color

string

default: #ffffff

platform_settings.widget.text_color

string

default: #000000

platform_settings.widget.btn_color

string

default: #000000

platform_settings.widget.btn_text_color

string

default: #ffffff

platform_settings.widget.border_color

string

default: #e1e1e1

platform_settings.widget.focus_color

string

default: #000000

platform_settings.widget.border_radius

integer

platform_settings.widget.btn_radius

integer

platform_settings.widget.action_text

string

platform_settings.widget.start_call_text

string

platform_settings.widget.end_call_text

string

platform_settings.widget.expand_text

string

platform_settings.widget.listening_text

string

platform_settings.widget.speaking_text

string

platform_settings.widget.shareable_page_text

string

platform_settings.data_collection

object

Show child attributes

platform_settings.data_collection.{key}

object

Show child attributes

platform_settings.data_collection.{key}.type

enum<string>

required

Available options:

boolean,

string,

integer,

number

platform_settings.data_collection.{key}.description

string

required

platform_settings.overrides

object

Show child attributes

platform_settings.overrides.conversation_config_override

object

Show child attributes

platform_settings.overrides.conversation_config_override.agent

object

Show child attributes

platform_settings.overrides.conversation_config_override.agent.prompt

object

Show child attributes

platform_settings.overrides.conversation_config_override.agent.first_message

boolean

default: false

platform_settings.overrides.conversation_config_override.agent.language

boolean

default: false

platform_settings.overrides.conversation_config_override.tts

object

Show child attributes

platform_settings.overrides.conversation_config_override.tts.voice_id

boolean

default: false

platform_settings.overrides.custom_llm_extra_body

boolean

default: false

platform_settings.ban

object

Show child attributes

platform_settings.ban.at_unix

integer

required

platform_settings.ban.reason_type

enum<string>

required

An enumeration.

Available options:

safety,

manual

platform_settings.ban.reason

string

platform_settings.safety

object

Safety object that has the information of safety evaluations based on used voice.

Show child attributes

platform_settings.safety.ivc

object

Safety evaluation of the agent. Prompt and first message is taken into account. The unsafe reason is provided from the evaluation

Show child attributes

platform_settings.safety.ivc.is_unsafe

boolean

default: false

platform_settings.safety.ivc.llm_reason

string

default:

platform_settings.safety.ivc.safety_prompt_version

integer

default: 0

platform_settings.safety.ivc.matched_rule_id

enum<string>[]

Available options:

sexual_minors,

forget_moderation,

extremism,

scam_fraud,

political,

self_harm,

illegal_distribution_medical,

sexual_adults,

unknown

platform_settings.safety.non_ivc

object

Safety evaluation of the agent. Prompt and first message is taken into account. The unsafe reason is provided from the evaluation

Show child attributes

platform_settings.safety.non_ivc.is_unsafe

boolean

default: false

platform_settings.safety.non_ivc.llm_reason

string

default:

platform_settings.safety.non_ivc.safety_prompt_version

integer

default: 0

platform_settings.safety.non_ivc.matched_rule_id

enum<string>[]

Available options:

sexual_minors,

forget_moderation,

extremism,

scam_fraud,

political,

self_harm,

illegal_distribution_medical,

sexual_adults,

unknown

platform_settings.privacy

object

Show child attributes

platform_settings.privacy.record_voice

boolean

default: true