Create a new AI agent with specified configuration including system prompt, tools, and model settings
API Key authentication for programmatic access. Include your API key in the Authorization header as: Bearer your_api_key_here
Request body for creating a new AI agent with configuration
Name of the agent
Description of what the agent does
System prompt that defines the agent behavior and instructions
Array of tool names that this agent can use
Input configuration defining what inputs the agent expects
Model alias must be a valid model from the registry
Optional context documents to provide to the agent
Optional reference to the Tines story import template used to create this agent
Optional planning behavior; defaults to auto
never, always, auto Optional structured output schema definition for this agent
Successful response
The newly created agent with generated ID and timestamps
Unique identifier for the agent
Unique identifier of the organization that owns this agent
Human-readable name of the agent
Description of what the agent does and its purpose
Timestamp when the agent was created
Timestamp when the agent was last updated
Timestamp when the agent was soft-deleted, or null if not deleted
Array of tool names that this agent can use during execution
Array of input configurations defining what inputs the agent expects
Model alias specifying which LLM model to use for this agent
Planning behavior mode for the agent
auto, never, always Type of agent: response (standard) or detection (security detection)
response, detection Whether this agent is built in to the system (true for built-in (system-managed) agents)
Type of built-in agent (e.g., threat-validation, threat-hunt)
threat-relevancy, threat-research Optional array of context documents (e.g., Google Docs, Notion pages) to provide to the agent
Optional reference to the Tines story import template used to create this agent
Optional array of triggers configured for this agent
Optional JSON schema describing the structured output the agent should emit (may include x-cotool metadata for preset UX; metadata is stripped before sending to the LLM)
Category of detection (only for detection agents)
baseline_anomaly, first_occurrence, sequence_pattern, privilege_escalation, custom, The SIEM query for the detection (only for detection agents)
The SIEM platform type (e.g., splunk, elastic) for the detection (only for detection agents)
How often the detection runs: hourly, daily, or weekly (only for detection agents)
hourly, daily, weekly, Whether this detection is a suggestion (only for detection agents)
Timestamp of when this detection was last run (only for detection agents)