---
name: datap-list-a-dataset
description: Use when the person wants to sell or publish a dataset on datap. Drafts the listing, sets fields and price, and hands over the few steps only a person can do, each with a link.
---

# List a dataset on datap

Keep replies short. Do each agent step with a tool; for each person step give one line and the link. `datatap_get_listing_checklist` shows what is done and what is next at any point.

## Before you start
Confirm the person has the right to sell the data. Only data the person has the right to sell goes further. Leave out, without opening more than you need to recognise it: personal data about identifiable people (contacts, customers, patients, messages, location history), credentials and keys, an employer's or client's confidential data, and anything bought or downloaded under a licence that forbids redistribution. If ownership is unclear, ask the person; do not guess.

## Steps
1. **Seller terms (person, once per workspace).** If `datatap_create_dataset_draft` answers that the terms must be accepted, send the person to https://datap.com/provider/datasets/new and continue when they are back.
2. **Draft (agent).** `datatap_create_dataset_draft`: a specific name buyers would search for, a description that says what a row is, coverage, time span and how it is collected, a category, the licence text, and the mode: `append_only` for events that never change (sales, posts, readings), `entity_state` for things that are updated in place (companies, products, profiles).
3. **Fields (agent).** `datatap_set_dataset_schema`: every field with its type, one stable primary key, and up to 20 public fields for the free sample. Never make a sensitive field public.
4. **Price (agent).** `datatap_suggest_price` returns a recommended model and amounts with comparable listings. Show the person the recommendation in one or two lines, then `datatap_set_dataset_price` once they agree.
5. **Ingestion key (person).** Created at https://datap.com/provider/datasets/DATASET_ID/api-keys and shown once. Ask the person to store it as DATATAP_INGEST_KEY in the environment that will upload. Do not ask them to paste it into the conversation.
6. **First upload (script).** Follow https://datap.com/docs/ingestion (Python and JavaScript examples). CSV, NDJSON, JSON or Parquet. Wait for the ingestion job to succeed; `datatap_get_my_dataset` shows it.
7. **Update schedule (agent).** If the data will be refreshed, `datatap_set_update_schedule` with an interval the person can reliably meet. See the `keep-dataset-updated` guide.
8. **Sharing permissions (person).** Which fields buyers get, permitted uses and retention: https://datap.com/provider/datasets/DATASET_ID
9. **Payouts (person, paid listings only).** Stripe onboarding at https://datap.com/provider/earnings
10. **Publish (person).** The Publish button on https://datap.com/provider/datasets/DATASET_ID

## Pricing in one paragraph
Three models. **snapshot**: one price for the whole dataset as it stands; best for data that will not change. **usage**: a price per block of rows the buyer actually takes; best for large datasets people filter. **subscription**: a monthly fee that includes a number of rows, then a per-row rate; best for data refreshed on a schedule. Updates to rows a buyer already owns are either included or charged at a separate update rate. Amounts are integers in minor units (pence or cents). Buyers never pay twice for the same row.

These steps use the datap connector (https://datap.com/mcp/private). Setup: https://datap.com/docs/agents
