Analog documentation
Analog lets any AI understand websites as code. 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.
Each structured result preserves the page as ordered sections of labeled records, prose, and navigation. It is saved locally, so you can reopen it and keep working without fetching the page again. No site-specific schema or API is required.
Choose a mode
Section titled “Choose a mode”Analog offers two modes. Choose based on what you need from the page.
Structured extraction
Section titled “Structured extraction”Use structured extraction when a page contains items you need to filter, compare, export, or revisit:
analog get https://quotes.toscrape.com/js/Analog renders the page in a browser on your machine, sends the page content to Analog, and returns its records, prose, and navigation in document order. The result is saved locally under a handle for querying, exporting, reopening, and comparing later.
Local Markdown conversion
Section titled “Local Markdown conversion”Use local Markdown when the page is mainly prose, or when you need to read it once rather than work with its records:
analog get https://example.com/article --mode localLocal mode fetches the page and converts it to Markdown entirely on your machine. It sends nothing to Analog and does not require an account.
Not sure which mode fits? Assess the page before extracting. For the complete data boundary, see Privacy.
Start with your task
Section titled “Start with your task”- Run Analog for the first time: Quickstart takes you from installation to a saved result.
- Capture more of a page: Fetching covers pagination, load-more buttons, and collapsed details.
- Work with a saved result: Work with results shows how to inspect, query, export, reopen, and compare what came back.
- Use Analog from an AI client: MCP server connects Analog to Claude, Cursor, and other MCP-compatible agents.
Looking for exact syntax? Use the CLI reference or Python API reference. Known boundaries and common failures live in Limitations and Troubleshooting.
If you are an AI agent setting yourself up, start with /agent-onboarding.md.