Skip to main content

Updates an existing assistant agent with version state

PUT 
/api/v2/agent/assistant/:agentCode/:agentVersionState

Updates an existing assistant agent with version state

Request​

Path Parameters

    agentCode stringrequired

    The code of the conversational agent to update.

    agentVersionState stringrequired

    The version state of the new agent version.

Query Parameters

    culture string

    Use this param to override the culture of the response. Options: - en (default) - es

Body

The conversational agent update request containing details.

    general

    object

    required

    name stringrequired

    Possible values: non-empty and <= 64 characters

    description stringnullable

    Possible values: <= 2000 characters

    behaviour

    object

    required

    systemDefinition stringrequired

    Possible values: non-empty and <= 40000 characters

    initialMessage stringrequired

    Possible values: non-empty and <= 2000 characters

    conversationObjective stringnullable

    Possible values: <= 1000 characters

    conversationStarters string[]nullable

    model

    object

    required

    main

    circular(AIModelReq)

    id uuidnullable
    modelCode stringnullable

    The model code associated with the AI model to be used by the agent. Required if Id is not provided.

    vendor stringnullable

    The vendor associated with the AI model to be used by the agent. Required if Id is not provided.

    parameters

    object

    Request DTO for configuring AI model execution parameters.

    frequencyPenalty floatnullable

    A float value between -2 and 2 that penalizes the likelihood of an agent reusing similar completions.

    logitBIAS stringnullable

    A JSON object that maps tokens to an associated bias value from -100 to 100.

    maxTotalTokens int64nullable

    The maximum number of tokens shared between the Prompt and the Output.

    maxOutputTokens int32nullable

    The maximum number of tokens that can be generated by the AI in the response.

    presencePenalty floatnullable

    A float value between -2 and 2 that penalizes the likelihood of an agent reusing phrases already present in the input.

    reasoningEffort stringnullable

    The reasoning level of the agent's response (e.g., "low", "medium", "high").

    stopSequences string[]nullable

    Defines sequences of tokens upon which the agent should stop generating completions.

    temperature floatnullable

    A float value between 0 and 2 that determines the randomness of the agent's responses.

    topK floatnullable

    A float value that specifies the Top-K sampling parameter.

    topP floatnullable

    A float value between 0 and 1 that specifies the probability cutoff for generating the response (nucleus sampling).

    useVision boolean

    Defines if the agent should use Vision capabilities.

    usePluginEmulator booleannullable

    Defines if the agent should use Plugin Emulator instead of native tools.

    enableThinking boolean

    Defines if the agent should use extended thinking (Alibaba Cloud and DeepSeek only).

    knowledge

    object

    knowledgeSources

    object[]

    nullable

    List of files to be integrated

  • Array [

    • id uuidnullable
    currentVersionId uuidnullable
    fileId uuidnullable
    url stringnullable

    metadata

    object

    description stringnullable

    Possible values: <= 1000 characters

    tags stringnullable

    Possible values: <= 1021 characters

  • ]

    • datasetSources uuid[]nullable

    List of dataset ids to be integrated as knowledge sources

    advancedSettings

    object

    integrationMode stringnullable

    The mode of knowledge integration, which determines how the knowledge is integrated into the agent's context. If not provided, default values will be set

    embeddingModelId uuidrequired

    The model used to generate the embeddings IF changed, all files will be reprocessed.

    relevance doublenullable

    Possible values: <= 1

    The relevance score of the knowledge If not provided, default values will be set

    sectionsToRetrieve int32nullable

    Possible values: >= 1 and <= 100

    Sets a limit on the number of sections to retrieve If not provided, default values will be set

    skills

    object[]

    nullable

    List of skills to be added to the agent.

  • Array [

    • oneOf

    Request model for creating or updating an Agent Executor skill within an agent.

    type stringrequired

    Possible values: non-empty, [AgentExecutor]

    The skill type identifier. Always "AgentExecutor" for agent executor skills.

    settings

    object

    required

    agentCode stringrequired

    Possible values: <= 50 characters, Value must match regular expression ^[a-zA-Z0-9\-]*$

    The code of the agent to execute when this skill is invoked.

    agentVersion int32nullable

    The version of the agent to use for this skill.

    code stringrequired

    Possible values: <= 50 characters, Value must match regular expression ^[a-zA-Z0-9]+$

    The unique code for this skill configuration.

    description stringnullable

    Possible values: <= 4000 characters

    The description of this skill configuration.

    includeInAgentResult boolean

    When set to true, the skill result will be returned within the agent result.

    reuseOutput boolean

    When set to true, the skill results will be cached and reused throughout a conversation as long as the input parameters remain the same.

    active boolean

    Only active skills will be imported during agent execution.

  • ]

Responses​

Conversational agent updated successfully.

Loading...