Foresea Predict API

openapi: 3.2.0
info:
  title: Foresea Intelligence Predict API
  description: '## Overview


    The **Foresea Intelligence API** turns market questions, linked evidence, and

    prediction-market prices into probabilistic forecasts that can sit behind

    trading dashboards, alerting systems, and research workflows. It supports binary, multiple-choice, numeric, and date forecasts. Each response

    includes a structured **rationale**, typed forecast fields, optional evidence

    sources fetched from live news, and optional market-edge analysis when you pass

    a market-implied probability. ---


    ## Quick start


    ```bash

    curl -X POST https://foresea.ink/predict \

    -H "Content-Type: application/json" \

    -d ''{

    "question": "Will the Fed cut rates before September 30, 2026?",

    "question_type": "binary",

    "market_platform": "Polymarket",

    "market_probability": 0.54,

    "variant": "variant0_neutral_baseline"

    }''

    ```


    ---


    ## Question types


    Use `question_type` when your client already knows the shape of the question. If omitted, the model will try to infer it. | Type | Response shape |

    |---|---|

    | `binary` | `predicted_answer` is `Yes`/`No`; `confidence` is 0–1 |

    | `multiple_choice` | `options` contains per-option probabilities; `predicted_answer` is the top option |

    | `numeric` | `range_forecast` contains `p10`, `p50`, `p90`, and optional `unit` |

    | `date` | `range_forecast` contains date percentiles |


    For multiple-choice forecasts, pass `options` when the answer set is known. ---


    ## Prediction-market context


    Pass `market_probability` when you know the current market-implied probability

    for the target outcome. For binary markets this defaults to the `Yes` side; set

    `market_outcome` to compare against `No` or a named multiple-choice outcome. Foresea returns `market_analysis` with model probability, market probability,

    percentage-point edge, and a stance (`model_above_market`, `model_below_market`,

    `in_line`, or `not_comparable`). ---


    ## Trading execution


    Foresea can preview and submit guarded orders for Kalshi and Polymarket:


    | Endpoint | Purpose |

    |---|---|

    | `GET /trading/accounts` | Return configured/not-configured server-account status without exposing secrets |

    | `POST /trading/accounts/check` | Validate caller-supplied own-account (BYO) credentials without storing them |

    | `POST /trading/preview` | Normalize and validate an order without placing it |

    | `POST /trading/orders` | Submit a live order after explicit confirmation |


    Live trading is disabled unless `FORESEA_ENABLE_TRADING=true`. Every live order

    requires a signed-in user session, `execute=true`, and the exact confirmation

    phrase `PLACE REAL ORDER`. Market/IOC/FOK-style orders are separately blocked

    unless `FORESEA_ALLOW_MARKET_ORDERS=true`. Exchange credentials come from a shared server account (environment variables /

    Secret Manager mounts) or an authenticated user''s **encrypted account

    connection**. The connection endpoint accepts a user''s credentials once, validates

    them with a unique data-encryption key wrapped by Cloud KMS, and never returns

    them to the browser. Inline `venue_credentials` are rejected on public preview

    and order calls; encrypted own-account trading remains gated by

    `FORESEA_ENABLE_BYO_TRADING=true` (independent of `FORESEA_ENABLE_TRADING`). `/agent/analyze` may return a directional recommendation, but it never places an

    order. Use `/trading/preview` first, review the normalized order, then call

    `/trading/orders` only when you intend to submit a real order. ---


    ## Prompt variants


    The `variant` field controls how the LLM is prompted.'
  contact:
    name: Pareel Amre
    url: https://foresea.ink/
    email: pareel.amre@gmail.com
  license:
    name: MIT
    url: https://opensource.org/licenses/MIT
  version: 1.0.0
tags:
- name: Predict
paths:
  /v1/batch/forecast:
    post:
      tags:
      - Predict
      summary: Batch forecast multiple prediction market questions concurrently
      description: Execute concurrent calibrated forecasts across a list of market questions.
      operationId: batch_forecast_route_v1_batch_forecast_post
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/BatchForecastRequest'
        required: true
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                additionalProperties: true
                type: object
                title: Response Batch Forecast Route V1 Batch Forecast Post
        '422':
          description: Validation Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/HTTPValidationError'
components:
  schemas:
    HTTPValidationError:
      properties:
        detail:
          items:
            $ref: '#/components/schemas/ValidationError'
          type: array
          title: Detail
      type: object
      title: HTTPValidationError
    BatchForecastItem:
      properties:
        id:
          anyOf:
          - type: string
          - type: 'null'
          title: Id
        question:
          type: string
          maxLength: 800
          minLength: 1
          title: Question
        market_platform:
          anyOf:
          - type: string
          - type: 'null'
          title: Market Platform
        market_probability:
          anyOf:
          - type: number
          - type: 'null'
          title: Market Probability
        market_url:
          anyOf:
          - type: string
          - type: 'null'
          title: Market Url
      type: object
      required:
      - question
      title: BatchForecastItem
    BatchForecastRequest:
      properties:
        questions:
          items:
            $ref: '#/components/schemas/BatchForecastItem'
          type: array
          maxItems: 20
          minItems: 1
          title: Questions
        model:
          anyOf:
          - type: string
          - type: 'null'
          title: Model
      type: object
      required:
      - questions
      title: BatchForecastRequest
    ValidationError:
      properties:
        loc:
          items:
            anyOf:
            - type: string
            - type: integer
          type: array
          title: Location
        msg:
          type: string
          title: Message
        type:
          type: string
          title: Error Type
        input:
          title: Input
        ctx:
          type: object
          title: Context
      type: object
      required:
      - loc
      - msg
      - type
      title: ValidationError