# Web archive

The `archive` tool reads the Internet Archive's Wayback Machine, with Common Crawl as a second source. Every answer carries the capture date and a link to the capture.

| Action | What you get |
| --- | --- |
| `get` | A page as it was on a date (`at`: `2023`, `2023-06`, `2023-06-15`, an ISO time or `2 years ago`), read like a live page. |
| `snapshots` | Every capture, or only the captures where the content really changed (`every: change`). |
| `diff` | What changed between a date and today, or two dates: lines, sections and fields when you pass a schema. |
| `history` | A field or section across versions, for example every plan and its price since 2021. |
| `urls` | Every address the archive holds for a site or path, with first and last capture; `check_live` lists the ones that are gone. |

```json
{ "action": "history", "url": "https://www.example.com/pricing", "since": "2021", "prompt": "every plan and its monthly price" }
```

## Limits

- Only pages someone archived can be read.
- The Internet Archive limits heavy automated use. Skryp paces its requests and keeps captures it has read, so large jobs take longer rather than failing.
- When a capture is an empty JavaScript shell, Skryp rebuilds it in a browser from the archive's own copies of its scripts (2 credits).

Source: https://skryp.dev/docs/archive