# Update agent (/docs/api-reference/agents/updateAgent) > Update an existing agent. Send only the fields you want to change. **PUT** `/agents/{agent_id}` Update an existing agent. Send only the fields you want to change. ```yaml operationId: updateAgent requestBody: required: true content: application/json: schema: type: object description: Agent configuration. properties: name: type: string description: Name for the agent. example: Customer Support Agent welcome_message: type: string description: Initial message the agent will say when answering a call. example: Hello! How can I help you today? context_breakdown: type: array description: >- List of context breakdowns, each containing `title`, `body`, and optional `is_enabled`. items: type: object required: - title - body properties: title: type: string description: Title of the breakdown. example: Purpose body: type: string description: Body of the breakdown — the detailed prompt content. example: This agent helps customers with product inquiries and support issues. is_enabled: type: boolean default: true description: Whether this section is included in the prompt. call_type: type: string enum: - Incoming - Outgoing description: Call type of the assistant. transcriber: type: object description: Configuration for the speech-to-text transcriber. properties: provider: type: string enum: - deepgram_stream - cartesia - sarvam - azure_stream description: The speech-to-text provider to use. example: deepgram_stream model: type: string enum: - nova-3 - nova-2 description: The model to use for transcription (required when provider is `deepgram_stream`). example: nova-3 silence_timeout_ms: type: integer description: Silence timeout in milliseconds. example: 400 numerals: type: boolean description: Convert numbers from words to digits. punctuate: type: boolean description: Add punctuation to the transcript. smart_format: type: boolean description: Apply smart formatting to the transcript. diarize: type: boolean description: Identify different speakers in the transcript. model: type: object description: Configuration for the language model. properties: model: type: string enum: - azure-gpt-4.1-mini - azure-gpt-4.1-nano - azure-gpt-4o - azure-gpt-4o-mini - gemini-2.5-flash - gemini-2.5-flash-lite - gpt-3.5-turbo - gpt-4.1-mini - gpt-4.1-nano - gpt-4o - gpt-4o-mini - gpt-5.1 - llama-3.3-70b-versatile description: The language model to use. example: gpt-4o-mini temperature: type: number minimum: 0 maximum: 1 description: Controls randomness in the model's output (0.0 to 1.0). example: 0.7 voice: type: object description: Configuration for the text-to-speech voice. properties: provider: type: string enum: - eleven_labs - deepgram - google - cartesia - rime description: The voice provider to use. example: eleven_labs voice_id: type: string description: The specific external voice identifier from the provider. example: JBFqnCBsd6RMkjVDRZzb web_search: type: object description: Configuration for web search capabilities. properties: enabled: type: boolean description: Enable or disable web search functionality. provider: type: string enum: - DuckDuckGo description: The search provider to use. example: DuckDuckGo post_call_actions: type: object description: Side effects that fire once the call ends. Configure email, webhook, or both. properties: email: type: object properties: enabled: type: boolean recipients: type: array description: Email addresses that should receive the notification. items: type: string format: email example: - support@example.com include: type: array description: Which sections to include in the email body. items: type: string enum: - summary - extracted_variables - fullConversation - sentiment extracted_variables: type: array description: Variables the model should pull out of the conversation for the email. items: type: object required: - key - prompt properties: key: type: string description: Unique identifier for the variable in the post-call payload. example: customer_issue prompt: type: string description: Instruction for the model on what to pull out of the conversation. example: Identify the main issue the customer is experiencing. webhook: type: object properties: enabled: type: boolean url: type: string format: uri description: Endpoint that receives a POST with the call payload. example: https://your-webhook-endpoint.com/omnidim-callback include: type: array description: Which sections to include in the webhook body. items: type: string enum: - summary - extracted_variables - fullConversation - sentiment extracted_variables: type: array description: Variables the model should pull out of the conversation for the webhook. items: type: object required: - key - prompt properties: key: type: string description: Unique identifier for the variable in the post-call payload. example: customer_issue prompt: type: string description: Instruction for the model on what to pull out of the conversation. example: Identify the main issue the customer is experiencing. transfer: type: object description: Conditional call transfer to a human agent or another number. properties: enabled: type: boolean transfer_options: type: array description: >- Where to transfer the call and under what condition. The first matching condition wins. items: type: object required: - number - transfer_condition - transfer_message properties: number: type: string description: Primary phone number to transfer to. Include country code with leading `+`. example: '+15551234567' type: type: string enum: - static - dynamic default: static description: | `static` transfers to `number`. `dynamic` lets the agent pick a number at runtime based on the conversation. backup_numbers: type: array description: Fallback numbers tried if the primary is unreachable. items: type: string transfer_condition: type: string description: Natural-language condition that triggers this transfer option. example: Transfer if the customer asks to speak with a human. transfer_message: type: string description: Message the agent says to the caller before executing the transfer. example: Please hold while I connect you to one of our agents. end_call: type: object description: Hang up automatically when a condition is met. properties: enabled: type: boolean condition: type: string description: >- Natural-language condition that triggers ending the call. Only evaluated when `enabled` is true. example: End the call once the customer's issue is resolved. message: type: string description: What the agent says before hanging up. example: Thank you for contacting us. Have a great day! background_track: type: object description: Ambient background noise that plays under the agent's voice. properties: enabled: type: boolean description: Whether to mix the ambient track under the agent's audio. name: type: string enum: - call_center - filler - office - office_1 - restaurant description: Ambient track to mix under the agent. volume: type: number minimum: 0 maximum: 1 default: 0.2 description: Volume level on a 0–1 scale. Default 0.2. initial_ringing_sound_enabled: type: boolean description: Plays a ringing tone after the call is picked up, until the agent starts speaking. voicemail: type: object description: Voicemail / answering-machine handling for outbound calls. properties: enabled: type: boolean description: Detect voicemail and react instead of speaking to a machine. message: type: string description: Message to leave when voicemail is detected. languages: type: array description: Languages the agent should support. Pass each language as a display-name string. items: type: string enum: - English - English (India) - English (US) - Hindi - Bengali - Spanish - Tamil - Marathi - Telugu - Gujarati - French example: - English - Hindi responses: '200': description: Updated agent. content: application/json: schema: type: object properties: id: type: integer description: Agent identifier. name: type: string status: type: string description: >- Build state of the agent after the update. Always `Completed` once the operation returns. example: id: 6365 name: Updated Support Agent status: Completed ``` **Python SDK** ```python from omnidimension import Client client = Client(api_key) # Update an existing agent agent_id = "your_agent_id_here" update_data = { "name": "Updated Support Agent", "welcome_message": "Hello! I'm your updated support assistant. How can I help you today?", "model": { "temperature": 0.8 } } response = client.agent.update(agent_id, update_data) print(response) ``` **curl** ```bash curl -X PUT "https://backend.omnidim.io/api/v1/agents/{agent_id}" \ -H "Authorization: Bearer $OMNIDIM_API_KEY" ```