> ## Documentation Index > Fetch the complete documentation index at: https://docs.together.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create a dedicated endpoint > Creates a new dedicated endpoint for serving models. The endpoint starts automatically after creation. You can deploy any supported model on hardware configurations that meet the model's requirements. ## OpenAPI ````yaml POST /endpoints openapi: 3.1.0 info: title: Together APIs description: The Together REST API. See https://docs.together.ai for more details. version: 2.0.0 termsOfService: https://www.together.ai/terms-of-service contact: name: Together Support url: https://www.together.ai/contact license: name: MIT url: https://github.com/togethercomputer/openapi/blob/main/LICENSE servers: - url: https://api.together.ai/v1 security: - bearerAuth: [] paths: /endpoints: post: tags: - Endpoints summary: Create a dedicated endpoint description: >- Creates a new dedicated endpoint for serving models. The endpoint starts automatically after creation. You can deploy any supported model on hardware configurations that meet the model's requirements. operationId: createEndpoint requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/CreateEndpointRequest' responses: '200': description: '200' content: application/json: schema: $ref: '#/components/schemas/DedicatedEndpoint' '403': description: Unauthorized content: application/json: schema: $ref: '#/components/schemas/ErrorData' '500': description: Internal error content: application/json: schema: $ref: '#/components/schemas/ErrorData' x-codeSamples: - lang: Python label: Together AI SDK (v2) source: | # Docs for v1 can be found by changing the above selector ^ from together import Together import os client = Together( api_key=os.environ.get("TOGETHER_API_KEY"), ) endpoint = client.endpoints.create( model="Qwen/Qwen3.5-9B-FP8", hardware="1x_nvidia_a100_80gb_sxm", autoscaling={ "min_replicas": 2, "max_replicas": 5, } ) print(endpoint.id) - lang: Python label: Together AI SDK (v1) source: | from together import Together import os client = Together( api_key=os.environ.get("TOGETHER_API_KEY"), ) endpoint = client.endpoints.create( model="Qwen/Qwen3.5-9B-FP8", hardware="1x_nvidia_a100_80gb_sxm", min_replicas=2, max_replicas=5, ) print(endpoint.id) - lang: TypeScript label: Together AI SDK (TypeScript) source: | import Together from "together-ai"; const client = new Together({ apiKey: process.env.TOGETHER_API_KEY, }); const endpoint = await client.endpoints.create({ model: "Qwen/Qwen3.5-9B-FP8", hardware: "1x_nvidia_a100_80gb_sxm", autoscaling: { max_replicas: 5, min_replicas: 2, } }); console.log(endpoint.id); - lang: JavaScript label: Together AI SDK (JavaScript) source: | import Together from "together-ai"; const client = new Together({ apiKey: process.env.TOGETHER_API_KEY, }); const endpoint = await client.endpoints.create({ model: "Qwen/Qwen3.5-9B-FP8", hardware: "1x_nvidia_a100_80gb_sxm", autoscaling: { max_replicas: 5, min_replicas: 2, } }); console.log(endpoint.id); - lang: Shell label: cURL source: | curl -X POST "https://api.together.ai/v1/endpoints" \ -H "Authorization: Bearer $TOGETHER_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "Qwen/Qwen3.5-9B-FP8", "hardware": "1x_nvidia_a100_80gb_sxm", "autoscaling": { "max_replicas": 5, "min_replicas": 2 } }' components: schemas: CreateEndpointRequest: type: object required: - model - hardware - autoscaling properties: display_name: type: string description: A human-readable name for the endpoint example: My Llama3 70b endpoint model: type: string description: The model to deploy on this endpoint example: deepseek-ai/DeepSeek-R1 hardware: type: string description: The hardware configuration to use for this endpoint example: 1x_nvidia_a100_80gb_sxm autoscaling: $ref: '#/components/schemas/Autoscaling' description: Configuration for automatic scaling of the endpoint disable_prompt_cache: deprecated: true type: boolean description: This parameter is deprecated and no longer has any effect. default: false disable_speculative_decoding: type: boolean description: Whether to disable speculative decoding for this endpoint default: false state: type: string description: The desired state of the endpoint enum: - STARTED - STOPPED default: STARTED example: STARTED inactive_timeout: type: integer description: >- The number of minutes of inactivity after which the endpoint stops automatically. Set to null, omit, or set to 0 to disable automatic timeout. nullable: true example: 60 availability_zone: type: string description: >- Create the endpoint in a specified availability zone (e.g., us-central-4b) DedicatedEndpoint: type: object description: Details about a dedicated endpoint deployment required: - object - id - name - display_name - model - hardware - type - owner - state - autoscaling - created_at properties: object: description: The object type, which is always `endpoint`. const: endpoint id: type: string description: Unique identifier for the endpoint example: endpoint-d23901de-ef8f-44bf-b3e7-de9c1ca8f2d7 name: type: string description: System name for the endpoint example: devuser/deepseek-ai/DeepSeek-R1-a32b82a1 display_name: type: string description: Human-readable name for the endpoint example: My DeepSeek R1 endpoint model: type: string description: The model deployed on this endpoint example: deepseek-ai/DeepSeek-R1 hardware: type: string description: The hardware configuration used for this endpoint example: 8x_nvidia_h200_140gb_sxm type: type: string enum: - dedicated description: The type of endpoint example: dedicated owner: type: string description: The owner of this endpoint example: devuser state: type: string enum: - PENDING - STARTING - STARTED - STOPPING - STOPPED - ERROR description: Current state of the endpoint example: STARTED autoscaling: $ref: '#/components/schemas/Autoscaling' description: Configuration for automatic scaling of the endpoint created_at: type: string format: date-time description: Timestamp when the endpoint was created example: '2025-02-04T10:43:55.405Z' ErrorData: type: object required: - error properties: error: type: object properties: message: type: string nullable: false type: type: string nullable: false param: type: string nullable: true default: null code: type: string nullable: true default: null required: - type - message Autoscaling: type: object description: Configuration for automatic scaling of replicas based on demand. required: - min_replicas - max_replicas properties: min_replicas: type: integer format: int32 description: >- The minimum number of replicas to maintain, even when there is no load examples: - 2 max_replicas: type: integer format: int32 description: The maximum number of replicas to scale up to under load examples: - 5 securitySchemes: bearerAuth: type: http scheme: bearer x-bearer-format: bearer x-default: default ````