> ## Documentation Index > Fetch the complete documentation index at: https://docs.tokz.dev/llms.txt > Use this file to discover all available pages before exploring further. # Answer typed decisions about texts > Asks one or more typed questions (choice, multilabel, bool, score) about up to 64 texts and returns, per text and question, the best option, the selected options, a probability distribution over the options, and for scores the expected value. This runs an OpenDecision model: the text is sent to a Tokz-run model service. It is not stored or logged; only counters are kept. Billed per token at the price of the chosen model, from the prepaid credit balance, and only when the request succeeds. Give the name of a fine tune your organization owns as model and leave out decisions to run that model on the API server's own CPU, up to 16 texts per request. The answer is then under the key decision. ## OpenAPI ````yaml /openapi.json post /decide openapi: 3.1.0 info: title: tokz API version: 1.0.0 description: >- Typed decisions about text from the open OpenDecision models. Text sent to /decide is read by the model and never stored or logged; only counters are kept. Fine tunes are built in the dashboard, which is not part of this API. contact: name: tokz url: https://tokz.dev servers: - url: https://api.tokz.dev description: Production - url: http://127.0.0.1:3999 description: Local dev (dev:pglite) security: - bearerAuth: [] tags: - name: Decisions description: Typed decisions about text, answered by the OpenDecision models. - name: Fine tunes description: >- The files of a model you trained. Starting, canceling and deleting one happen in the dashboard. - name: Operations description: Health and diagnostics. paths: /decide: post: tags: - Decisions summary: Answer typed decisions about texts description: >- Asks one or more typed questions (choice, multilabel, bool, score) about up to 64 texts and returns, per text and question, the best option, the selected options, a probability distribution over the options, and for scores the expected value. This runs an OpenDecision model: the text is sent to a Tokz-run model service. It is not stored or logged; only counters are kept. Billed per token at the price of the chosen model, from the prepaid credit balance, and only when the request succeeds. Give the name of a fine tune your organization owns as model and leave out decisions to run that model on the API server's own CPU, up to 16 texts per request. The answer is then under the key decision. operationId: decide requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/DecideRequest' examples: supportTicket: summary: Route a support ticket and flag urgency value: inputs: >- I was charged twice for my May invoice and the app keeps logging me out. decisions: queue: type: choice question: Which team should handle this ticket? options: - billing - technical - account - sales urgent: type: bool fineTune: summary: Call a fine tune by name value: model: support-triage inputs: I was charged twice for my May invoice. responses: '200': description: >- One object per input, keyed by decision name. A fine tune uses the fixed key decision. content: application/json: schema: $ref: '#/components/schemas/DecideResponse' '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '402': $ref: '#/components/responses/InsufficientCredits' '413': description: >- The request body exceeds 4 MiB, or (input_too_large) a text needs more memory than a hosted fine tune with this many labels allows. Nothing was charged. content: application/json: schema: $ref: '#/components/schemas/Error' '429': description: >- Too many requests for this API key, or (decide_org_busy) too many texts in progress on hosted fine tunes for this organization. Nothing was charged. headers: Retry-After: description: >- Seconds to wait before sending again. Present on most refusals of a fine tune. schema: type: integer content: application/json: schema: $ref: '#/components/schemas/Error' '502': description: The model service did not answer. Nothing was charged; retry. content: application/json: schema: $ref: '#/components/schemas/Error' '503': description: >- The model service or hosted fine tunes are busy (decide_busy), a fine tune is loading or cannot load (model_unavailable), Fine Tune is off (fine_tune_off), or the model service is not available on this deployment (decide_unavailable). Nothing was charged. headers: Retry-After: description: >- Seconds to wait before sending again. Present on most refusals of a fine tune. schema: type: integer content: application/json: schema: $ref: '#/components/schemas/Error' components: schemas: DecideRequest: type: object required: - inputs additionalProperties: false properties: model: type: string description: >- A catalog model (opendecision-small, opendecision-large-packed, opendecision-large), or the name of a fine tune your organization owns. default: opendecision-large-packed inputs: oneOf: - type: string minLength: 1 maxLength: 20000 - type: array minItems: 1 maxItems: 64 items: type: string minLength: 1 maxLength: 20000 description: >- One text or up to 64, each at most 20,000 characters. A fine tune takes up to 16. decisions: type: object minProperties: 1 maxProperties: 16 additionalProperties: $ref: '#/components/schemas/DecideDecision' description: >- Questions keyed by name, 1 to 16 per request. Required for a catalog model, and refused for a fine tune, which already knows its question and labels. DecideResponse: type: object required: - id - model - results - usage properties: id: type: string model: type: string results: type: array items: type: object additionalProperties: $ref: '#/components/schemas/DecideResult' description: >- One object per input, in order, keyed by decision name. A fine tune uses the fixed key decision. usage: type: object required: - billed_tokens - cost_usd properties: billed_tokens: type: integer cost_usd: type: number description: billed_tokens at the model's list price. Error: type: object required: - error properties: error: type: string description: A code or a short message. Never contains payload content. message: type: string description: >- What went wrong and what to do next, when the API has a sentence for it. requestId: type: string DecideDecision: type: object required: - type additionalProperties: false description: >- One question. choice and multilabel need options (2 to 256, unique ids); bool and score take none. properties: type: type: string enum: - choice - multilabel - bool - score question: type: string maxLength: 1000 options: type: array minItems: 2 maxItems: 256 items: $ref: '#/components/schemas/DecideOption' threshold: type: number exclusiveMinimum: 0 exclusiveMaximum: 1 default: 0.5 description: >- multilabel only: an option is selected when its probability reaches this. low: type: integer default: 0 description: score only. high: type: integer default: 5 description: score only. step: type: integer minimum: 1 default: 1 description: score only. The range must hold 2 to 256 values. DecideResult: type: object required: - type - best - selected - policy - value properties: type: type: string enum: - choice - multilabel - bool - score best: type: string description: The highest-probability option. selected: type: array items: type: string description: >- multilabel: every option whose value reaches the threshold. Otherwise the best option. policy: type: object additionalProperties: type: number description: Distribution over the options; sums to 1. value: type: object additionalProperties: type: number description: Per-option probability that the option is correct. score: type: - number - 'null' description: 'score only: the expected value.' DecideOption: oneOf: - type: string minLength: 1 maxLength: 500 - type: object required: - id additionalProperties: false properties: id: type: string minLength: 1 maxLength: 200 name: type: - string - 'null' maxLength: 500 description: type: - string - 'null' maxLength: 2000 responses: BadRequest: description: The request body is malformed or a field is out of range. content: application/json: schema: $ref: '#/components/schemas/Error' Unauthorized: description: Missing, malformed, revoked, or incorrect API key. content: application/json: schema: $ref: '#/components/schemas/Error' InsufficientCredits: description: >- The account behind the key has no remaining prepaid credits. Add credits and retry; the request was not processed. content: application/json: schema: type: object properties: error: type: string enum: - insufficient_credits balance: type: integer description: Remaining balance in cents (0 when exhausted). required: - error - balance securitySchemes: bearerAuth: type: http scheme: bearer description: >- Your API key: `Authorization: Bearer tokz_live__`. Only a SHA-256 hash of the secret is stored. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.