> ## Documentation Index
> Fetch the complete documentation index at: https://docs.veri.studio/llms.txt
> Use this file to discover all available pages before exploring further.

# Preview dataset rows



## OpenAPI

````yaml /api-reference/openapi.json get /v1/datasets/{dataset_id}/rows
openapi: 3.1.0
info:
  title: Veri API
  description: >-
    REST API for the Veri RL post-training platform. All requests require a
    Bearer API key (`vk_` prefix).
  license:
    name: ''
  version: 0.1.0
servers:
  - url: https://api.veri.studio
    description: Production
security: []
tags:
  - name: Training jobs
    description: Create, monitor, and manage training jobs.
  - name: Datasets
    description: Upload and connect training datasets.
  - name: Deployments
    description: Serve trained models and run inference.
  - name: Volumes
    description: Persistent file storage mounted into jobs.
  - name: Models
    description: Custom model registry deployments serve from.
  - name: Regions
    description: Discover available launch regions.
  - name: GPU
    description: Live GPU availability by provider and region.
  - name: Code artifacts
    description: Upload custom training script bundles.
  - name: Billing
    description: Credit balance and transaction history.
  - name: API keys
    description: Create and revoke API keys.
  - name: Account
    description: The authenticated caller's identity.
  - name: Settings
    description: Account-level integrations (Weights & Biases).
  - name: Metrics
    description: Prometheus metrics export for your own observability stack.
  - name: Evaluators
    description: 'Evaluators: versioned scoring rules (LLM judge, code, human).'
  - name: Experiments
    description: >-
      Experiments: offline runs of a pinned dataset snapshot through a target,
      scored by pinned evaluators.
  - name: Annotation queues
    description: >-
      Human review: queue traces, threads and experiment items, reserve one at a
      time, score them and feed corrections back into datasets.
  - name: Monitors
    description: >-
      Monitors: evaluators scoring a sampled share of a deployment's live
      traffic.
  - name: Observability
    description: Agent conversations, agents and their traffic.
  - name: Public runs
    description: Unauthenticated reads of runs their owners published to Explore.
paths:
  /v1/datasets/{dataset_id}/rows:
    get:
      tags:
        - Datasets
      summary: Preview dataset rows
      operationId: rows
      parameters:
        - name: dataset_id
          in: path
          description: Dataset ID
          required: true
          schema:
            type: string
        - name: cursor
          in: query
          description: >-
            Byte-offset cursor from the previous page's `next_cursor`. Omit for
            the

            first page.
          required: false
          schema:
            type: integer
            format: int64
            minimum: 0
        - name: limit
          in: query
          description: Rows per page, 1..=100 (default 50).
          required: false
          schema:
            type: integer
            format: int32
        - name: start_row
          in: query
          description: >-
            0-based row number to start the page at (jump-to-row). Takes
            precedence

            over `cursor`. The first jump into a dataset builds a sparse row
            index

            (one full streaming read); every later jump is two small S3 reads.
          required: false
          schema:
            type: integer
            format: int64
            minimum: 0
        - name: snapshot
          in: query
          description: |-
            Stream view: read the dataset AS OF this snapshot (`snap-N`, `N`, or
            the full `<id>@snap-N`). Supersessions/tombstones up to its head are
            resolved. Omit for the live head. Switches the response to
            `{row_id, row}` objects with `next_cursor` = the last row_id.
          required: false
          schema:
            type: string
        - name: since
          in: query
          description: |-
            Stream view: only rows with row_id > since (pass the previous page's
            `next_cursor`). Implies the stream view even without `snapshot`.
          required: false
          schema:
            type: integer
            format: int64
      responses:
        '200':
          description: One page of raw dataset rows
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DatasetRowsResponse'
        '404':
          description: Not found or not owned by the caller
        '409':
          description: >-
            Connected dataset has no stored snapshot yet (available after its
            first training run)
      security:
        - bearerAuth: []
components:
  schemas:
    DatasetRowsResponse:
      type: object
      description: >-
        One page of dataset rows. Uploads: `rows` are the raw JSONL objects and

        `next_cursor` is an opaque byte offset into the stored JSONL; pass it
        back

        verbatim as `cursor` to fetch the next page (null after the last row).

        Stream view (stream datasets, or `snapshot`/`since` given): `rows` are

        `{row_id, row}` objects and `next_cursor` is the last row_id, to pass
        back

        as `since`. `total_rows` is set for uploads only.
      required:
        - object
        - rows
      properties:
        object:
          type: string
        rows:
          type: array
          items: {}
        next_cursor:
          type:
            - integer
            - 'null'
          format: int64
          minimum: 0
        total_rows:
          type:
            - integer
            - 'null'
          format: int32
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: API key
      description: API key with the `vk_` prefix. Create one from the dashboard.

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.