# Analog documentation

Source: https://getanalog.io/docs/

Learn what Analog does, choose how to use a page, and find your next step.

**Analog is a web interface built for AIs.** It converts webpages into
the things they contain — products, articles, events, and other records
— so AIs can query, compare, export, and use them directly. It works
without prior examples or a site-specific schema or API.

Each structured result preserves the page as ordered sections of
labeled records, prose, and navigation. Use this view to get your
bearings, find relevant items, and inspect their details. Results are
saved locally, so you can return to them without fetching the page again.

Keep using your agent's browser to navigate and interact. Supply its captured
page to `analog view` to search, inspect, and save the content. If your
environment cannot supply a page file, `analog browse` acquires a URL with
Analog's built-in browser. The [Quickstart](https://getanalog.io/docs/quickstart/) demonstrates
the Codex GUI capture route and the browsing fallback.

## Choose how to use Analog

| Use Analog from | Start here |
| --- | --- |
| A terminal or a coding agent that can run commands | [Quickstart](https://getanalog.io/docs/quickstart/) or the [agent setup instructions](https://getanalog.io/agent-onboarding.md) |
| A Python program | [Install in your project](https://getanalog.io/docs/install/#install-in-a-python-project), then [run the Python example](https://getanalog.io/docs/extract/#in-python) |
| An app that supports local MCP servers | [MCP setup](https://getanalog.io/docs/mcp/) |

## Choose a mode

Analog offers two modes. Choose based on what you need from the page.

### Structured view

View the page's sections, records, and surrounding text together.
The structure helps you find and understand its content; the same
records are available for filtering, comparing, and exporting:

After [capturing a page](https://getanalog.io/docs/quickstart/#3-view-a-page):

```bash example=illustrative fragment=true
analog view ./about.capture.json --save-as about --find "Frankie"
```

Analog sends the captured page content and URL for extraction, saves the
result locally, and returns matching evidence with its handle and capture
qualifications. Query, export, reopen, or compare the save later. This mode
requires an account. Use `analog view about` for an overview.

### Local Markdown conversion

Use local Markdown when reading the page's prose gives you what you
need:

```bash example=illustrative fragment=true
analog view ./about.capture.json --mode local --markdown
```

Local mode converts the supplied page to Markdown entirely on your machine.
It sends nothing to Analog and does not require an account. `browse` also
supports local mode when you need Analog to fetch the page.

For help choosing, [assess the page](https://getanalog.io/docs/assess/). The complete data
boundary is in [Privacy](https://getanalog.io/docs/privacy/).

## Start with your task

- **Capture more of a page:** [Fetching](https://getanalog.io/docs/fetching/) covers
  pagination, load-more buttons, and collapsed details.
- **Work with a saved result:** [Working with results](https://getanalog.io/docs/results/)
  shows how to inspect, query, export, reopen, and compare what came
  back.
- **Convert HTML to Markdown in Python:** [Marcato](https://getanalog.io/docs/marcato/)
  turns an HTML string into Markdown with one rendering policy.

Looking for exact syntax? Use the [CLI reference](https://getanalog.io/docs/cli-reference/)
or [Python API reference](https://getanalog.io/docs/api-reference/). Known boundaries and
common failures live in [Limitations](https://getanalog.io/docs/limitations/) and
[Troubleshooting](https://getanalog.io/docs/troubleshooting/).