# Parsli

> Parsli extracts structured data from documents — invoices, receipts, statements, forms, scans, photographs, and handwriting. Define what you want once; every document of that kind comes back in the same shape. No fixed layouts, no per-sender templates.

Any documentation URL below also serves raw markdown: append `.md`, or send `Accept: text/markdown`.

## Start here

- [Parsli documentation](https://parsli.co/docs.md): Turn any document into structured data. Start with the five-minute walkthrough, or jump straight to the API.
- [What Parsli does](https://parsli.co/docs/start-here/what-parsli-does.md): Parsli reads documents and returns structured data — no templates to draw, no fixed layouts, no rules to maintain.
- [Your first extraction](https://parsli.co/docs/start-here/your-first-extraction.md): Create a parser, send it one real document, and get structured data back. About five minutes, no card required.
- [How Parsli works](https://parsli.co/docs/start-here/how-parsli-works.md): The object model — parsers, documents, fields, inbound, outbound — and how a document moves through it.

## Get documents in

- [Get documents in](https://parsli.co/docs/get-documents-in.md): Upload, email, or send from another system. Every route into a parser, and which one fits your situation.
- [By email](https://parsli.co/docs/get-documents-in/by-email.md): Every parser has its own address. Forward documents to it, or connect a Gmail mailbox and let Parsli watch for them.
- [From another system](https://parsli.co/docs/get-documents-in/from-another-system.md): Send documents programmatically — with an API key from code you control, or via the inbound webhook from a no-code tool.

## Define what to extract

- [Define what to extract](https://parsli.co/docs/define-what-to-extract.md): Fields are how you describe the data you want. Naming them well is the single biggest driver of accuracy.
- [Tables and nested data](https://parsli.co/docs/define-what-to-extract/tables-and-nested-data.md): Line items, transactions, and grouped values. How to model repeating data so it comes back as rows instead of a mess.
- [Extraction instructions](https://parsli.co/docs/define-what-to-extract/extraction-instructions.md): When a field name is not enough. Per-field and parser-wide instructions, with examples that work and ones that do not.

## Check the results

- [Check the results](https://parsli.co/docs/check-the-results.md): Reading extracted output, understanding confidence and empty values, and fixing an extraction that is wrong.

## Send data out

- [Send data out](https://parsli.co/docs/send-data-out.md): Every place extracted data can land — spreadsheets, accounting systems, automation tools, and your own endpoints.
- [Spreadsheets and documents](https://parsli.co/docs/send-data-out/spreadsheets-and-documents.md): A Google Sheet that fills itself with one formula, a Google Doc per extraction, or your own XLSX template populated automatically.
- [Accounting systems](https://parsli.co/docs/send-data-out/accounting-systems.md): File extracted bills and invoices straight into QuickBooks Online, Xero, or Zoho Books.
- [Automation tools](https://parsli.co/docs/send-data-out/automation-tools.md): Zapier, Make, and Power Automate — reaching everything Parsli does not connect to directly.
- [Webhooks](https://parsli.co/docs/send-data-out/webhooks.md): Send every extracted result to your own endpoint. Configuration, payload, authentication, and the two failure modes that matter.

## Build with the API

- [Build with the API](https://parsli.co/docs/api.md): Send documents from your own system and get structured JSON back. Two routes — inline for small files, upload-and-poll for everything else.
- [Authentication](https://parsli.co/docs/api/authentication.md): API keys are bearer tokens scoped to one parser, shown once at creation. How to create, use, scope, and rotate them.
- [Extract a document](https://parsli.co/docs/api/extract-a-document.md): POST a base64 file, get structured JSON back in the same response. The inline route, its limits, and how results are shaped.
- [Large and slow files](https://parsli.co/docs/api/large-and-slow-files.md): The three-step upload-and-poll route. No size ceiling, no request timeout — the only reliable way to handle documents you did not choose.
- [Receiving results](https://parsli.co/docs/api/receiving-results.md): Have Parsli call you when a document finishes, instead of asking repeatedly. The webhook payload in full.
- [Errors](https://parsli.co/docs/api/errors.md): Every status code the API returns, what causes it, and whether retrying will help.

## Credits and billing

- [How pages are counted](https://parsli.co/docs/billing/how-pages-are-counted.md): One page is one credit. What counts as a page, what happens on failure, and how to predict a bill.
- [Free pages and plans](https://parsli.co/docs/billing/free-pages-and-plans.md): You get free pages on signup with no card. They expire on a clock that starts at your first extraction — here is exactly how that works.

## Teams

- [Teams](https://parsli.co/docs/teams.md): Share parsers and one pool of credits across a group, with roles that control who can change what.

## Premium workspaces

- [Premium workspaces](https://parsli.co/docs/premium.md): Complete products for a whole business process, built on top of extraction and set up with you. Not something you switch on yourself.

## Dictionary

- [Dictionary](https://parsli.co/docs/dictionary.md): Every term Parsli uses, defined twice — once in plain English, once precisely. Including the words we used to use.

## Machine-readable resources

- [Full documentation](https://parsli.co/llms-full.txt): every page concatenated, for a single-request load
- [Dictionary as JSON](https://parsli.co/docs/dictionary.json): the complete object model — every term with a plain definition, a precise definition, aliases, and relationships