> ## Documentation Index > Fetch the complete documentation index at: https://assemblyai.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Create a decision > Answers a list of typed questions about one input. Each answer carries a probability instead of generated text. Available on the US server only. See [Decisions](https://www.assemblyai.com/docs/llm-gateway/decisions) for the full guide. ## OpenAPI ````yaml api-reference/specs/llm-gateway.yaml POST /decisions openapi: 3.1.0 info: title: AAI Chat Completions API description: API for generating text with various language models. version: 1.0.0 servers: - url: https://llm-gateway.assemblyai.com/v1 description: Production Server x-fern-server-name: Production security: - ApiKey: [] paths: /decisions: post: summary: Create a decision description: >- Answers a list of typed questions about one input. Each answer carries a probability instead of generated text. Available on the US server only. See [Decisions](https://www.assemblyai.com/docs/llm-gateway/decisions) for the full guide. operationId: createDecision requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/DecisionRequest' examples: decision_example: summary: Check a support call value: model: gpt-6-luna input: >- Agent: Thanks for calling Northwind Internet, this is Priya. Can I get your account number? Customer: Sure, it's 4471-2209. Agent: Thanks. And can you confirm the address on the account? Customer: 18 Harbor Lane. My internet has dropped three times today and I work from home. Agent: I'm sorry about that. There's an outage in your area. A technician is already on it, and it should be fixed by 5 PM. Customer: Okay, that's fine. Thanks for letting me know. questions: - type: predicate name: confirmed_account_details instructions: >- Did the agent ask the caller to confirm their account details? - type: choice name: call_outcome instructions: How did the call end? choices: - value: resolved description: The issue was fixed or fully answered on the call - value: follow_up description: The customer was told the issue will be fixed later - value: escalated description: The call was transferred or escalated - type: score name: customer_sentiment instructions: How did the customer feel at the end of the call? levels: - label: negative description: Upset or frustrated - label: neutral description: Neither upset nor pleased - label: positive description: Satisfied or thankful responses: '200': description: One answer per question, in request order. content: application/json: schema: $ref: '#/components/schemas/DecisionResponse' '400': description: >- The request is invalid, or the model's provider needs an opt-in your account hasn't accepted. `metadata.errors` lists each problem. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' default: description: An unexpected error occurred. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' components: schemas: DecisionRequest: type: object description: >- The request body for the decisions endpoint. Unknown fields are rejected. properties: model: type: string enum: - gpt-6-luna - jev-1.13.0 description: >- The model to use. `jev-1.13.0` requires accepting the Jev opt-in on the [Data Controls](https://www.assemblyai.com/dashboard/settings/data-controls) page. input: description: >- The evidence every question is asked about. A string, or an array of user messages made of `input_text` parts. oneOf: - type: string - type: array items: type: object additionalProperties: true questions: type: array minItems: 1 maxItems: 400 items: $ref: '#/components/schemas/DecisionQuestion' description: The questions to answer. Names must be unique. model_region: type: string enum: - global description: >- Optional. Set to `global` for global routing. Supported on `gpt-6-luna` only. prompt_cache_key: type: string description: Optional. Requests with the same key share a prompt cache. fallbacks: type: array items: $ref: '#/components/schemas/DecisionFallback' description: >- Models to try in order if the main request fails. See [Specify fallback models](https://www.assemblyai.com/docs/llm-gateway/fallback). fallback_config: $ref: '#/components/schemas/FallbackConfig' required: - model - input - questions DecisionResponse: type: object properties: request_id: type: string format: uuid model: type: string description: The model that answered. answers: type: array items: $ref: '#/components/schemas/DecisionAnswer' description: One answer per question, in request order. usage: $ref: '#/components/schemas/DecisionUsage' ErrorResponse: type: object properties: code: type: integer format: int32 description: HTTP status code for the error. message: type: string description: A human-readable description of the error. request_id: type: string format: uuid description: Unique identifier for the request. metadata: type: object description: >- Optional. Present on 400 responses with per-field validation details. properties: errors: type: array items: type: string description: List of specific validation failure messages. required: - code - message - request_id DecisionQuestion: type: object properties: type: type: string enum: - predicate - choice - score description: >- `predicate` asks whether something is true, `choice` picks one of `choices`, and `score` places the input on the scale given by `levels`. name: type: string description: >- A name for the question, unique within the request. The answer carries the same name. instructions: description: >- The question. A string, or an array of `input_text` parts (and `input_image` parts on `gpt-6-luna`). oneOf: - type: string - type: array items: type: object additionalProperties: true choices: type: array minItems: 2 items: $ref: '#/components/schemas/DecisionChoice' description: >- Required for `choice` questions, and not allowed for other types. Values must be unique. levels: type: array minItems: 2 maxItems: 10 items: $ref: '#/components/schemas/DecisionLevel' description: >- Required for `score` questions, and not allowed for other types. Ordered from lowest to highest. required: - type - name - instructions DecisionFallback: type: object description: >- A fallback for a decisions request. `input` and `questions` are inherited from the main request when omitted. Nested fallbacks aren't supported. properties: model: type: string enum: - gpt-6-luna - jev-1.13.0 input: description: Overrides the main request's `input`. oneOf: - type: string - type: array items: type: object additionalProperties: true questions: type: array items: $ref: '#/components/schemas/DecisionQuestion' description: Overrides the main request's `questions`. model_region: type: string enum: - global description: Overrides the main request's `model_region`. required: - model FallbackConfig: type: object description: >- Configuration for fallback behavior. See [Specify fallback models](https://www.assemblyai.com/docs/llm-gateway/fallback) for more details. properties: retry: type: boolean default: true description: >- Whether to automatically retry the request once after 500ms on failure. Defaults to `true`. depth: type: integer minimum: 1 maximum: 2 default: 1 description: >- The maximum number of fallbacks to traverse. Defaults to `1`, with a maximum of `2`. DecisionAnswer: type: object properties: type: type: string enum: - predicate - choice - score name: type: string description: The question's name. probability: type: number description: '`predicate` only. The probability that the statement is true.' choice: type: string description: '`choice` only. The most likely option''s value.' score: type: number description: '`score` only. The expected level index, weighted by probability.' probabilities: type: array items: $ref: '#/components/schemas/DecisionProbability' description: >- `choice` and `score` only. The probability of each option or level, in request order. confidence: type: number description: >- `choice` and `score` only. How concentrated the probabilities are, from 0 to 1. DecisionUsage: type: object properties: input_tokens: type: integer input_tokens_details: type: object properties: cached_tokens: type: integer cache_write_tokens: type: integer output_tokens: type: integer output_tokens_details: type: object properties: reasoning_tokens: type: integer total_tokens: type: integer DecisionChoice: type: object properties: value: type: string description: The value returned when this option is chosen. description: type: string description: What this option means. required: - value DecisionLevel: type: object properties: label: type: string description: The level's name. description: type: string description: What this level means. required: - label DecisionProbability: type: object properties: value: description: >- The option's value for `choice` questions, or the level's index for `score` questions. oneOf: - type: string - type: integer label: type: string description: '`score` only. The level''s label.' probability: type: number securitySchemes: ApiKey: type: apiKey in: header name: Authorization ````