---
title: "Credits and billing"
url: https://www.datawebot.com/docs/billing
description: "How extraction Jobs use credits: a price per page fetched, holds and settlement, free failed Jobs, trial credits and what happens when your balance runs low."
author: datawebot team
---

Find documentation[Documentation home](https://www.datawebot.com/docs)

Concepts

[Products, variants and fields](https://www.datawebot.com/docs/data)

Developers

[Connect your agent](https://www.datawebot.com/docs/agents) · [API quickstart](https://www.datawebot.com/docs/api/quickstart) · [API reference](https://www.datawebot.com/docs/api/reference) · [Run datawebot on a schedule](https://www.datawebot.com/docs/schedule) · [Where datawebot is listed](https://www.datawebot.com/docs/directory)

Account

[Credits and billing](https://www.datawebot.com/docs/billing) · [Troubleshooting](https://www.datawebot.com/docs/troubleshooting)

Machine-readable

[OpenAPI specification ↗](https://www.datawebot.com/openapi.json) · [Documentation index ↗](https://www.datawebot.com/llms.txt)

Account

# Credits and billing

How extraction Jobs use credits: a price per page fetched, holds and settlement, free failed Jobs, trial credits and what happens when your balance runs low.

## Credits pay for pages fetched

Extraction Jobs are paid in credits from your account wallet, never charged directly to a card. Each endpoint has a price per page fetched. By default that is 5 credits per search page, 5 per catalogue page and 1 credit per product page. Rows, products and variants are output measures, not billing units.

Your agent sees each endpoint's current price in find_endpoints before it submits a Job; GET /v1/job-blueprints shows it as price for scripts. Some endpoints, such as newly qualified sites, can have a different price or none yet; the listed price is the one that applies.

## Searches and product reads

search_products and get_product (and POST /v1/requests) run Jobs for you and charge once per request. A search costs at most the dearer of the shop's search endpoint and shopping search, 5 credits at default prices; a product read costs the shop's product page price, 1 credit by default. check_site (GET /v1/sites) gives a shop's prices before anything runs, and get_account shows credits left; both are free.

Pass max_credits to set a budget: if a request could cost more, nothing runs and nothing is charged (price_exceeds_max). For a search on a shop with its own endpoint, a failed read is free and shopping search answers instead; if the shop's page loads and has nothing for the query, the request ends there and pays for that page only.

## Holds and settlement

When a Job is submitted, the credits for the pages it asks for (max_pages, up to the endpoint's limit; 1 page by default) are set aside. When the Job finishes it pays only for pages fetched successfully, and the rest of the hold is released. Asking for 10 pages and fetching 5 costs 5 pages.

The price is fixed when the Job is submitted. A later price change does not affect a Job that is already running.

## Failed Jobs are free; empty Jobs pay for pages that loaded

A Job that fails, including on a blocked page, costs no credits; its hold is released. A Job that loads its pages but finds nothing pays for the pages the worker confirmed loaded, and nothing when it can't confirm them. Repeating a submission with the same idempotency key returns the same Job and is not charged again.

## Trial credits and low balance

Every account starts with 200 free trial credits. When your balance cannot cover a new Job's hold, the Job is refused with insufficient_credits (HTTP 402). Ask for fewer pages to lower the hold.

Buy more credits on the dashboard under Billing: Scout $10 for 10,000 credits, Agent $25 for 32,500, Fleet $100 for 160,000 and Swarm $600 for 1,200,000. You pay on Stripe's checkout page and the credits are added once the payment is confirmed. Searches and product reads through MCP use the same credits: 5 per search and 1 per product read, and failed ones are free.

## Earlier store extraction API

Runs from the earlier store extraction API (/api/product/runs) keep their settled charges. Its page budgets and one-credit-per-page rules do not apply to extraction Jobs. Contact support with a run or Job ID to investigate a charge.

[← Products, variants and fields](https://www.datawebot.com/docs/data) · [API quickstart →](https://www.datawebot.com/docs/api/quickstart)

Need help? [Contact support](https://www.datawebot.com/support).