# Web

URL: /code/platform/web · Id: `capability:web`

> Web searches and reads the open web for agents and keeps each page read as a Source with its fetch time.

Web searches and reads the open web for agents and keeps each page read as a [Source](/business/nouns/source) with its fetch time.

## Contract
Every page an [Agent](/business/nouns/agent) reads becomes a Source, so any claim built on it can be traced and refetched.

| Rule | Detail |
| --- | --- |
| Search | Returns ranked results with title, URL and snippet |
| Read | Fetches a page as clean text, with its fetch time |
| Source | Each read is stored as a Source with URL, content hash and time |
| Robots | Reads honor the site's crawl rules |
| [Cost](/business/nouns/cost) | Each search and read is counted against the caller's [Limit](/business/nouns/limit) |

## Verbs
| Verb | Object | Event |
| --- | --- | --- |
| [Source.create](/business/nouns/source/create) | [Source](/business/nouns/source) | `Source.created` |

## Interfaces
- [API](/code/api/reference)
- [SDK](/code/sdk/reference)
- [CLI](/code/cli/reference)
- [MCP](/code/mcp/reference)

## Example
A grant-writing Agent searches for open arts grants in one state and reads six funder pages. Each page is stored as a Source with its fetch time, and the draft cites them by Source.