> ## Documentation Index > Fetch the complete documentation index at: https://docs.perplexity.ai/llms.txt > Use this file to discover all available pages before exploring further. # Answer questions > Evaluate one or more named questions against the content in `state`. Mix `noul`, `choice`, and `score` questions in one request; the response returns one answer per question under the same name, plus the model name and token usage. Set `model` to `decider-27b`; a missing or unknown model returns `400`. ## OpenAPI ````yaml openapi-gateway-preview.json post /decisions openapi: 3.0.3 info: title: Decisions API description: >- Ask yes/no, multiple-choice, and rating questions about your content and get probabilities back. Input tokens are billed at $0.04 per million; output tokens are free. Authenticate with your Perplexity API key as `Authorization: Bearer `. version: 0.1.0 servers: - url: https://api.perplexity.ai/v1 security: [] paths: /decisions: post: summary: Answer questions description: >- Evaluate one or more named questions against the content in `state`. Mix `noul`, `choice`, and `score` questions in one request; the response returns one answer per question under the same name, plus the model name and token usage. Set `model` to `decider-27b`; a missing or unknown model returns `400`. operationId: decisions_v1_decisions_post requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/DecisionsRequest' examples: mixed-questions: summary: Three question types about one product review value: model: decider-27b state: title: Battery died after two weeks review: >- The headphones sound great, but the battery stopped charging after two weeks. questions: defect: type: noul instructions: Does the review report a product defect? sentiment: type: choice instructions: What is the overall sentiment of the review? criteria: positive: Mostly satisfied mixed: Praise and complaints in one review negative: Mostly dissatisfied severity: type: score instructions: How severe is the reported problem? criteria: - Cosmetic - Inconvenient - Product unusable responses: '200': headers: x-request-id: $ref: '#/components/headers/x-request-id' x-ratelimit-limit: $ref: '#/components/headers/x-ratelimit-limit' x-ratelimit-remaining: $ref: '#/components/headers/x-ratelimit-remaining' x-ratelimit-used: $ref: '#/components/headers/x-ratelimit-used' x-ratelimit-reset: $ref: '#/components/headers/x-ratelimit-reset' content: application/json: schema: $ref: '#/components/schemas/DecisionsResponse' examples: answers: summary: One answer per question, typed like its question value: model: decider-27b answers: defect: type: noul noul: 0.9424522889347015 sentiment: type: choice choice: mixed confidence: 0.9255246944002182 probabilities: positive: 0.020649883775315993 mixed: 0.9503497962668123 negative: 0.02900031995787183 severity: type: score score: 1.7838686319784252 confidence: 0.7838686319784252 legend: '0': Cosmetic '1': Inconvenient '2': Product unusable probabilities: '0': 0.008423954913615923 '1': 0.199283458194343 '2': 0.7922925868920411 usage: input_tokens: 367 output_tokens: 3 description: One answer per question, under the question's name. '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '404': $ref: '#/components/responses/NotFound' '405': $ref: '#/components/responses/MethodNotAllowed' '413': $ref: '#/components/responses/PayloadTooLarge' '429': $ref: '#/components/responses/TooManyRequests' '500': $ref: '#/components/responses/InternalServerError' '502': $ref: '#/components/responses/BadGateway' '503': $ref: '#/components/responses/ServiceUnavailable' '504': $ref: '#/components/responses/GatewayTimeout' security: - HTTPBearer: [] components: schemas: DecisionsRequest: type: object title: DecisionsRequest properties: model: type: string title: Model description: >- The model to use: `decider-27b`, or `decider-27b-v0` to pin the current version. A missing or unknown model returns `400`. example: decider-27b state: anyOf: - type: string - type: object additionalProperties: true - type: array items: {} title: State description: >- The content every question in the request refers to: a string, an object, or an array. Images go in an array as OpenAI-style `image_url` parts with base64 PNG, JPEG, or WebP data URLs. example: >- The headphones sound great, but the battery stopped charging after two weeks. questions: type: object additionalProperties: $ref: '#/components/schemas/Question' minProperties: 1 title: Questions description: >- Questions keyed by a non-empty name you choose. The response uses the same names. 1 to 128 questions. example: defect: type: noul instructions: Does the review report a product defect? maxProperties: 128 required: - model - state - questions description: >- The content to evaluate and the named questions to ask about it. Unknown fields return `400`. Input must stay under 262,144 tokens and the body under 32 MiB. additionalProperties: false DecisionsResponse: type: object title: DecisionsResponse properties: model: type: string title: Model description: The model name you sent, echoed back. example: decider-27b answers: type: object additionalProperties: $ref: '#/components/schemas/Answer' minProperties: 1 title: Answers description: One answer per question, under the question's name. example: defect: type: noul noul: 0.9424522889347015 usage: allOf: - $ref: '#/components/schemas/Usage' description: Token counts for this request. example: input_tokens: 367 output_tokens: 3 required: - model - answers - usage description: >- Answers keyed by question name, plus the model that answered and token usage. Question: oneOf: - $ref: '#/components/schemas/NoulQuestion' - $ref: '#/components/schemas/ChoiceQuestion' - $ref: '#/components/schemas/ScoreQuestion' discriminator: propertyName: type mapping: choice: $ref: '#/components/schemas/ChoiceQuestion' noul: $ref: '#/components/schemas/NoulQuestion' score: $ref: '#/components/schemas/ScoreQuestion' description: A named question about `state`. Discriminated by `type`. Answer: oneOf: - $ref: '#/components/schemas/NoulAnswer' - $ref: '#/components/schemas/ScoreAnswer' - $ref: '#/components/schemas/ChoiceAnswer' discriminator: propertyName: type mapping: choice: $ref: '#/components/schemas/ChoiceAnswer' noul: $ref: '#/components/schemas/NoulAnswer' score: $ref: '#/components/schemas/ScoreAnswer' description: One answer. Its `type` matches the question's `type`. Usage: type: object title: Usage properties: input_tokens: type: integer title: Input Tokens description: Input tokens processed for the request. example: 367 output_tokens: type: integer title: Output Tokens description: Output tokens generated for the answers. example: 3 required: - input_tokens - output_tokens description: >- Token counts for the request. Input tokens are billed at $0.04 per million; output tokens are free. ErrorResponse: type: object title: ErrorResponse properties: error: $ref: '#/components/schemas/Error' required: - error description: The body of an error response. NoulQuestion: type: object title: NoulQuestion properties: type: type: string enum: - noul title: Type description: Always `noul`. example: noul instructions: anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: Instructions description: >- The question to answer or the statement to check. Required unless `criteria` is set. example: Does the review report a product defect? criteria: allOf: - $ref: '#/components/schemas/NoulCriteria' description: >- Optional definitions of yes and no. Required unless `instructions` is set. example: 'true': Something is broken or not working. 'false': Normal wear or personal preference. required: - type description: >- A yes/no question, or a statement to check, answered with the probability of yes or true. ChoiceQuestion: type: object title: ChoiceQuestion properties: type: type: string enum: - choice title: Type description: Always `choice`. example: choice instructions: anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: Instructions description: What to decide when choosing between the options. example: What is the overall sentiment of the review? criteria: type: object additionalProperties: anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: Criteria description: >- Option names mapped to a description of when each applies. Use `null` to let the name speak for itself. 1 to 255 options. example: positive: Mostly satisfied mixed: Praise and complaints in one review negative: Mostly dissatisfied minProperties: 1 maxProperties: 255 required: - criteria - type description: A question that picks one of the options you define. ScoreQuestion: type: object title: ScoreQuestion properties: type: type: string enum: - score title: Type description: Always `score`. example: score instructions: anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: Instructions description: What to rate. example: How severe is the reported problem? criteria: type: array minItems: 1 items: anyOf: - type: string - type: object additionalProperties: true - type: array items: {} title: Criteria description: >- Ordered level descriptions. The index in this array is the level's score, starting at 0. 1 to 10 levels; use at least two. example: - Cosmetic - Inconvenient - Product unusable maxItems: 10 required: - criteria - type description: A question that rates the content on an ordered rubric. NoulAnswer: type: object title: NoulAnswer properties: type: type: string enum: - noul title: Type description: Always `noul`. example: noul noul: type: number title: Noul description: >- Probability of yes or true, from 0 to 1. Values near 0.5 mean the model is unsure. example: 0.9424522889347015 required: - noul - type description: >- Answer to a `noul` question: the probability that the answer is yes or the statement is true. ScoreAnswer: type: object title: ScoreAnswer properties: type: type: string enum: - score title: Type description: Always `score`. example: score score: type: number title: Score description: >- Probability-weighted average of the level indices. Can fall between two levels. example: 1.7838686319784252 confidence: type: number title: Confidence description: Confidence in the score, from 0 to 1. Higher means more certainty. example: 0.7838686319784252 legend: type: object additionalProperties: anyOf: - type: string - type: object additionalProperties: true - type: array items: {} title: Legend description: >- Level index, as a string, mapped back to the `criteria` entry it stands for. example: '0': Cosmetic '1': Inconvenient '2': Product unusable probabilities: type: object additionalProperties: type: number title: Probabilities description: >- Probability of each level, keyed like `legend`. Values sum to approximately 1. example: '0': 0.008423954913615923 '1': 0.199283458194343 '2': 0.7922925868920411 required: - score - confidence - legend - probabilities - type description: >- Answer to a `score` question: the expected score, the rubric, a confidence value, and the probability of each level. ChoiceAnswer: type: object title: ChoiceAnswer properties: type: type: string enum: - choice title: Type description: Always `choice`. example: choice choice: type: string title: Choice description: The option with the highest probability. example: mixed confidence: type: number title: Confidence description: >- Confidence in the selected option, from 0 to 1. Higher means more certainty. example: 0.9255246944002182 probabilities: type: object additionalProperties: type: number title: Probabilities description: >- Probability of each option, keyed by option name. Values are between 0 and 1 and sum to approximately 1. example: positive: 0.020649883775315993 mixed: 0.9503497962668123 negative: 0.02900031995787183 required: - choice - confidence - probabilities - type description: >- Answer to a `choice` question: the selected option, a confidence value, and the probability of each option. Error: type: object title: Error properties: message: type: string description: What went wrong. example: Noul question must have criteria or instructions type: type: string description: >- The error category, for example `invalid_request_error`, `invalid_request`, `invalid_api_key`, or `too_many_requests`. example: invalid_request code: description: >- A status code as a string or a number, or `null`. Do not branch on it; use the HTTP status. anyOf: - type: string nullable: true - type: integer nullable: true param: type: string nullable: true description: Present on some errors; `null` when present. required: - message - type description: Details of the error. NoulCriteria: type: object nullable: true title: NoulCriteria properties: 'true': anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: 'True' description: What a yes answer means for this question. example: The review describes something broken or not working. 'false': anyOf: - type: string nullable: true - type: object additionalProperties: true nullable: true - type: array items: {} nullable: true title: 'False' description: What a no answer means for this question. example: The review describes normal wear or personal preference. description: Optional definitions of what counts as yes and what counts as no. headers: x-request-id: description: A UUID that identifies the request. Quote it in support requests. schema: type: string format: uuid example: 7a1504a6-a884-48e6-af6e-0c3140aeb6db x-ratelimit-limit: schema: type: integer description: Requests allowed per second for your organization. x-ratelimit-remaining: schema: type: integer description: Requests left in the current second. x-ratelimit-used: schema: type: integer description: Requests used in the current second. x-ratelimit-reset: schema: type: integer description: When the window resets, in Unix seconds. Retry-After: schema: type: integer description: Seconds to wait before retrying. responses: BadRequest: description: >- The body is not valid JSON, `model` is missing or unknown, a field is unknown or the wrong type, a request limit is exceeded, or an image is not a supported data URL. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: message: >- Invalid model 'decider-27b-latest'. Permitted models can be found in the documentation at https://docs.perplexity.ai/docs/getting-started/models. type: invalid_request_error param: null code: null headers: x-request-id: $ref: '#/components/headers/x-request-id' Unauthorized: description: >- The API key is missing or invalid, or it was sent in `x-api-key`. No request id header. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: message: >- Invalid API key provided. You can find your API key at https://console.perplexity.ai. type: invalid_api_key code: 401 NotFound: description: Wrong path, including a trailing slash. Empty body, no request id. MethodNotAllowed: description: Wrong method. Empty body. headers: Allow: description: Always `POST`. schema: type: string example: POST x-request-id: $ref: '#/components/headers/x-request-id' PayloadTooLarge: description: The request body is over 32 MiB. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: null message: request body exceeds the maximum allowed size of 33554432 bytes param: null type: invalid_request_error headers: x-request-id: $ref: '#/components/headers/x-request-id' TooManyRequests: description: Over the request or token limit. Wait `Retry-After` seconds, then retry. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: null message: Request rate limit exceeded, please try again later. param: null type: too_many_requests headers: Retry-After: $ref: '#/components/headers/Retry-After' x-request-id: $ref: '#/components/headers/x-request-id' InternalServerError: description: The service failed. Retry with backoff. BadGateway: description: The model service was unreachable. Retry with backoff. ServiceUnavailable: description: The model service is temporarily unavailable. Retry with backoff. GatewayTimeout: description: >- The model did not answer in time, after about a minute. The body may be HTML rather than JSON. Retry, or send less input. securitySchemes: HTTPBearer: type: http scheme: bearer description: >- Your Perplexity API key, sent as `Authorization: Bearer `. The `x-api-key` header is not read. ````