---
title: "Ficha completa de un dataset (AI metadata)"
method: GET
path: "/v1/datasets/{id}"
tags: ["Datasets"]
---

# Ficha completa de un dataset (AI metadata)

`GET /v1/datasets/{id}`

## Request

- Base URL: `https://api.almanac.ar`
- URL: `https://api.almanac.ar/v1/datasets/{id}`
- Auth: none declared

Público, sin auth. Schema de columnas, ejemplos, gotchas, patrón de actualización, SLA de frescura, y `snapshot_date`/`row_count` del snapshot vigente. Equivalente a lo que el MCP server devuelve vía `get_dataset(id)`.

## Path parameters

- `id` string, required

## Response `200`

La ficha del dataset.

- DatasetDetailResponse
  - `dataset_id` string, required
  - `name` string, required
  - `description` string, nullable, required
  - `source` object, required
    - `id` string, required
    - `name` string, required
  - `frequency` string, nullable, required
  - `granularity` string, nullable, required
  - `historical_depth` string, nullable, required
  - `tags` string[], nullable, required
  - `access` 'included_in_pro', required
  - `country` string, nullable, required
  - `source_native_language` string, nullable, required
  - `update_pattern` 'incremental_append' | 'full_refresh' | 'scd2_snapshots' | 'null', nullable, required
  - `freshness_sla` DatasetFreshnessSla, nullable, required
    - `declared_frequency` 'daily' | 'weekly' | 'monthly' | 'quarterly' | 'annual' | 'irregular' | 'intraday', required
    - `max_acceptable_lag_hours` number, required
  - `content_advisory` string, nullable, required
  - `compliance_notes` DatasetComplianceNotes, nullable, required
    - `license` string, required
    - `attribution_required` string, nullable
    - `commercial_use` 'ok' | 'restricted' | 'prohibited', required
    - `notes` string, nullable
  - `snapshot_date` string, nullable, required: Fecha (YYYY-MM-DD) del snapshot vigente (`is_latest=true`), o null si el dataset todavía no tiene snapshot publicado. Si es una VIEW derivada, hereda la fecha del dataset padre.
  - `row_count` integer, nullable, required: Cantidad de filas del snapshot vigente, o null si no hay snapshot todavía.
  - `ai_metadata` DatasetAiMetadata, required
    - `llm_intro` string, nullable, required
    - `business_questions` string[], nullable, required
    - `example_queries` DatasetExampleQuery[], nullable, required
      - `question` string, required
      - `approach` string, required
      - `polars` string, nullable
      - `sql` string, nullable
    - `methodology_notes` string, nullable, required
    - `gotchas` DatasetGotcha[], nullable, required
      - `title` string, required
      - `detail` string, required
    - `related_datasets` DatasetRelated[], nullable, required
      - `id` string, required
      - `relation` string, required
      - `derived` boolean: true si ESTE dataset es una VIEW derivada del relacionado (hereda su snapshot).
    - `columns` DatasetColumn[], nullable, required
      - `name` string, required
      - `type` string, required
      - `description` string, required
      - `unit` string, nullable
      - `sample_value` unknown
      - `nullable` boolean
    - `summary_columns` string[], nullable, required
    - `explorable_columns` string[], nullable, required
    - `last_updated` string, nullable, required
  - `updated_at` string, required

## Other responses

- `404`: dataset_not_found — no existe, no está publicado, no alcanza el plan del caller (anónimo = tier free), o es admin-only. Mismo status/shape en los cuatro casos: no filtra cuál aplica.

---

[API](https://skmtc.dev/almanac/apis/almanac-api.md) · [All operations](https://skmtc.dev/almanac/apis/almanac-api/llms.txt) · [OpenAPI document](https://skmtc.dev/almanac/apis/almanac-api/revisions/f1b8ffd54263?raw)