> ## Documentation Index > Fetch the complete documentation index at: https://docs.trugen.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create Agent > Creates a new agent with avatar IDs, knowledge bases, and additional settings ## OpenAPI ````yaml POST /ext/agent openapi: 3.0.3 info: title: TruGen Developer Platform API version: 1.0.0 description: OpenAPI specification for MCP endpoint servers: - url: https://api.trugen.ai/v1 security: [] paths: /ext/agent: post: summary: Create an agent description: >- Creates a new agent with avatar IDs, knowledge bases, and additional settings requestBody: required: true content: application/json: schema: type: object properties: agent_name: type: string description: Display name for the avatar persona. default: Sofia - Friendly Support agent_system_prompt: type: string description: Behavioral prompt for the avatar interaction style. default: Speak in a warm, engaging tone and provide clear answers. avatars: type: array description: List of avatars associated with this agent. items: type: object properties: config: type: object description: >- Configuration for avatar behavior and runtime settings. properties: llm: type: object description: Large Language Model settings used by the avatar. properties: model: type: string description: Identifier of the LLM model. default: meta-llama/llama-4-maverick-17b-128e-instruct provider: type: string description: Provider used for LLM inference. default: groq url: type: string description: >- Custom provider URL of the LLM model. (Required if provider is 'custom') default: sample url token: type: string description: >- Custom provider API token for the LLM model. (Required if provider is 'custom') default: sample token stt: type: object description: Speech-to-Text model configuration. properties: model: type: string description: Speech recognition model identifier. default: flux-general-en provider: type: string description: STT provider service name. default: deepgram min_endpointing_delay: type: number description: Minimum endpointing delay in seconds. default: 0.3 max_endpointing_delay: type: number description: Maximum endpointing delay in seconds. default: 0.4 tts: type: object description: Text-to-Speech generation configuration. properties: model_id: type: string description: Voice synthesis model identifier. default: eleven_turbo_v2_5 provider: type: string description: TTS provider service name. default: elevenlabs voice_id: type: string description: Voice ID used to synthesize speech. default: ZUrEGyu8GFMwnHbvLhv2 turn_detector: type: boolean description: turn detector default: true avatar_key_id: type: string description: avatar_key_id default: 665a1170 idle_timeout: type: object description: idle_timeout configuration. properties: timeout: type: number description: timeout default: 30 filler_phrases: type: array description: filler_phrases items: type: string default: >- Hey it's been a while since we last spoke, are we still connected? welcome_message: type: object description: welcome_message configuration. properties: wait_time: type: number description: wait_time default: 2 messages: type: array description: messages items: type: string default: Hi, how are you doing today? warning_exit_message: type: object description: warning_exit_message configuration. properties: callout_before: type: number description: callout_before default: 10 messages: type: array description: messages items: type: string default: >- We are almost at the end of our call, thank you for your time. exit_message: type: object description: exit_message configuration. properties: max_call_duration: type: number description: max_call_duration default: 300 messages: type: array description: messages items: type: string default: >- We are at the end of our call, thank you for your time. exit_heads_up_message: type: object description: exit_heads_up_message configuration. properties: callout_before: type: number description: callout_before default: 10 messages: type: array description: messages items: type: string default: >- We are almost at the end of our call, thank you for your time. config: type: object description: Agent-level configuration properties. properties: timeout: type: integer description: >- Maximum time to wait for the user to join the call (in seconds). default: 240 memory: type: object description: Memory configuration for the agent. properties: isEnabled: type: boolean description: Whether memory is enabled for the agent. default: false instruction: type: string description: Instruction for how the agent should utilize memory. default: >- Sample instruction for the agent to utilize memory effectively in conversations. knowledge_base: type: array description: List of knowledge bases attached to this agent. nullable: true items: type: object properties: id: type: string format: uuid description: Knowledge base identifier. default: ac79226e-73e1-41fe-8cde-469ae4e244fa name: type: string description: Name of the knowledge base. default: Product Support Articles description: type: string description: Short summary of the knowledge base content. default: >- Contains FAQs and troubleshooting guides for product lines. tool: type: array description: List of tools attached to this agent. nullable: true items: type: object properties: id: type: string format: uuid description: Tool identifier. default: ac79226e-73e1-41fe-8cde-469ae4e244fa name: type: string description: Name of the tool. default: Product Support Articles mcp: type: array description: List of MCPs attached to this agent. nullable: true items: type: object properties: id: type: string format: uuid description: MCP identifier. default: ac79226e-73e1-41fe-8cde-469ae4e244fa name: type: string description: Name of the MCP . default: Product Support Articles record: type: boolean description: Determines whether voice call recordings should be stored. default: true callback_events: type: array description: Webhook events that should trigger callbacks. items: type: string default: utterance_committed example: - participant_left - agent.started_speaking - agent.stopped_speaking - agent.interrupted - user.started_speaking - user.stopped_speaking - utterance_committed - max_call_duration_timeout callback_url: type: string description: Webhook endpoint URL to send callback events to. example: https://webhooks.example.com/agent-events required: - avatars - agent_name - agent_system_prompt - config example: agent_name: Sample AI Agent agent_system_prompt: You're a helpful ai agent. Answer all questions appropriately. config: timeout: 240 memory: isEnabled: false instruction: sample memory instruction for the agent knowledge_base: - id: 4a0365e4-ced5-42f0-8933-b6880a0ce044 name: new kb 123 record: true callback_url: '' callback_events: - participant_left - max_call_duration_warning - max_call_duration_timeout - action_found avatars: - avatar_key_id: 665a1170 config: llm: model: meta-llama/llama-4-maverick-17b-128e-instruct provider: groq stt: model: flux-general-en provider: deepgram min_endpointing_delay: 0.3 max_endpointing_delay: 0.4 tts: model_id: eleven_turbo_v2_5 provider: elevenlabs voice_id: ZUrEGyu8GFMwnHbvLhv2 persona_name: Sample AI Agent persona_prompt: >- You're a helpful ai agent. Answer all questions appropriately. conversational_context: Sample Conversational Context idle_timeout: timeout: 30 filler_phrases: - >- Hey it's been a while since we last spoke, are we still connected? - >- I notice we haven't talked for a bit, is everything okay? welcome_message: wait_time: 2 messages: - Hi, how are you doing today? - Hello, how can I help you? warning_exit_message: callout_before: 10 messages: - >- We are almost at the end of our call, thank you for your time. - Thank you for your time. We will see you next time. exit_message: max_call_duration: 300 messages: - We are at the end of our call, thank you for your time. - Thank you for your time today. exit_heads_up_message: callout_before: 10 messages: - >- We are almost at the end of our call, thank you for your time. - Thank you for your time. We will see you next time. responses: '200': description: Agent created successfully! content: application/json: schema: type: object properties: id: type: string message: type: string example: id: 045302f0-9783-4ff2-846e-3888fb2a7894 message: Agent created successfully '400': description: Bad request content: application/json: schema: type: object properties: error: type: string example: error: Invalid request body '401': description: Unauthorized – missing or invalid API key content: application/json: schema: type: object properties: error: type: string example: error: 'Unauthorized: not a valid authorization api key' '500': description: Internal server error content: application/json: schema: type: object properties: error: type: string example: error: Error creating agent security: - apiKeyAuth: [] components: securitySchemes: apiKeyAuth: type: apiKey in: header name: x-api-key ````