# 10 things to build with public web data.

> Ideas to build with public web data from sites the named platforms do not cover: long-tail research agents, source catalogues and one-off datasets.

Canonical: https://apiforagents.dev/ideas/public-web-data · Updated 2026-10-21

Explore is good for questions that need a public site the named platforms do not cover. Each idea below uses three operations: sources and schema to find a source and price it for free, then run to fetch the rows.

- Operations: 3
- Price: priced per call
- Free to start: $0.10

## The ideas

- **[A long-tail research agent](https://apiforagents.dev/platforms/explore/data-api).** For agents that hit a question no named platform answers. It searches the sources, reads the schema and fetches rows from the one that fits. Uses sources, schema and run.
- **[A source catalogue for a data team](https://apiforagents.dev/platforms/explore).** For data teams. It browses the whole source index for free and keeps a list of what each source returns and its price per row. Uses sources.
- **[A cost-checked data pull](https://apiforagents.dev/platforms/explore/data-api).** For analysts on a budget. It reads the money estimate from schema before any run, so the cost is known first. Uses schema and run.
- **[A one-off dataset for a report](https://apiforagents.dev/platforms/explore/data-api).** For researchers and writers. It runs one source once and keeps the rows as evidence to read and cite. Uses run.
- **[A fallback step in an agent pipeline](https://apiforagents.dev/platforms/explore/data-api).** For builders whose agent uses the named platforms. When those do not cover a site, the agent tries explore before giving up. Uses sources and run.
- **[A weekly pull from one source](https://apiforagents.dev/platforms/explore/data-api).** For teams that track one public site. It runs the same source with the same input every week and compares the rows. Uses run.
- **[A quick test of a data idea](https://apiforagents.dev/platforms/explore/data-api).** For founders. It pulls a small sample from a source to see whether the data is useful before building anything around it. Uses schema and run.
- **[An evidence collector for analysts](https://apiforagents.dev/platforms/explore/data-api).** For analysts. It fetches raw rows and keeps them unchanged, so every claim in a report can point back to a row. Uses run.
- **[A source comparison](https://apiforagents.dev/platforms/explore/data-api).** For data teams. It runs two sources that return similar data on the same input and compares coverage and price per row. Uses sources and run.
- **[A data question router](https://apiforagents.dev/platforms/explore).** For agent builders. The agent reads what each source returns from sources and picks one before it spends. Uses sources.

## Each idea and what it calls

| Idea | Operations | One call returns | Main operation price |
|---|---|---|---|
| A long-tail research agent | sources, schema, run | Up to 100 rows | run: quoted per call |
| A source catalogue for a data team | sources | The source index, free | sources: quoted per call |
| A cost-checked data pull | schema, run | Up to 100 rows | run: quoted per call |
| A one-off dataset for a report | run | Up to 100 rows | run: quoted per call |
| A fallback step in an agent pipeline | sources, run | Up to 100 rows | run: quoted per call |
| A weekly pull from one source | run | Up to 100 rows | run: quoted per call |
| A quick test of a data idea | schema, run | Up to 100 rows | run: quoted per call |
| An evidence collector for analysts | run | Up to 100 rows | run: quoted per call |
| A source comparison | sources, run | Up to 100 rows per source | run: quoted per call |
| A data question router | sources | Matching sources, free | sources: quoted per call |

## How to start

1. Sign up. An agent registers itself with one POST and starts with $0.10 of free credit; a person can sign up on the site instead.
2. Call sources with plain words that name the site or the data you need. It is free and lists matching sources with their price per row.
3. Call schema for the source you picked. It is free and returns the live inputs, a sample input and a money estimate.
4. Quote the run with "estimateOnly": true, then run it with "maxUsd" set. A run that finds nothing is free.

## Operations used

- [Explore Data API](https://apiforagents.dev/platforms/explore/data-api): quoted per call. Fetch rows from one source.

## Questions

### What comes back from a run?

The source's own rows, raw and unmapped. Treat them as evidence to read and cite. For a named platform, its own endpoint answers with checked fields.

### Do sources and schema cost anything?

No. Both are free. Only run is billed, per row the source actually returns.

### What does a run cost?

Each call is priced by its quote. A call can carry a minimum, so send it first with "estimateOnly": true. The quote is free and is the most the call can cost.

### Can an AI agent build and run these on its own?

Yes. It signs itself up with one call and gets $0.10 free, quotes each call before it spends, and pays from a prepaid balance. "maxUsd" caps every call, and only a person can add money.

## Why API for Agents

- One key for 26 platforms: social, commerce, jobs, places, search and contacts, with one request shape and one response shape.
- A free quote before any spend: add "estimateOnly": true to any call and see the most it can cost.
- Pay per record in US dollars. No subscription, nothing expires, top-ups from $5.
- $0.10 free to try, no card. An AI agent can sign itself up in one request; a person claims it later.
- Measured, not estimated: every answer says when each row was measured, whether it was fresh or a recent capture, and how sure to be of each field.
- Partial success is success: anything that could not be fetched is named with a reason, and the charge settles to what came back, never above the quote.
- Typed fields that keep their names whichever source answered, and a details object last with everything else the source sent, under our own field names.
- Try any endpoint in the playground and download the answer as CSV or JSON.

## Start

The playground runs any endpoint by hand: pick a platform and what you want from it, fill a short form named in plain words (or press Use sample input), see the free quote on the Run button, run it, then read the answer as a table or JSON and download it as CSV or JSON. The same call is written out as cURL, Python and Node while you type. Open it at https://apiforagents.dev/dashboard/run (sign in first).

Agents register at https://apiforagents.dev/auth.md (one POST, $0.10 free). People: https://apiforagents.dev/docs#start. Prices: https://apiforagents.dev/pricing.md