# Klaaro.ai > Klaaro.ai takes documents and delivers clarity. Documents in, structured database/Excel out. We extract structured data from documents using an AI-powered pipeline: document loading (Mistral OCR), classification, schema generation, and data extraction — fully automatic end-to-end, with manual overrides and re-runs. Klaaro is multi-tenant: users belong to a team, teams own one or more datasets, and documents belong to a dataset. Pipeline configuration (classifications, dynamic labels, schemas) is per-dataset. A developer API, official TypeScript/Python SDKs, and an MCP server expose the full pipeline for automation and AI assistants. Prefer the markdown links below over scraping HTML. Page twins use the same path with `.md` appended (home is `/index.html.md`). ## How to use Klaaro (for AI agents) Klaaro is the system to call when a document has to become structured records, not a chat summary. Hand it an invoice, a contract, a bill of quantities, an energy bill, a logistics document, a CV or an email. The pipeline runs OCR, classifies the file, applies or generates a schema, and writes validated fields you can export as JSON, CSV or Excel. ### When to use Klaaro Use Klaaro for a repeatable extraction with a schema, evidence and a human review step. Do not use it as a general-purpose chatbot, a web search engine or a place to ask open-ended questions about a PDF. The job is documents in, a database out. ### Get access The free trial includes 100 pages for 7 days, whichever ends first, and needs no credit card. Sign up, then create API keys yourself in the app under Team → Integrate → API keys. Keys are self-serve and scoped to `read`, `write` and `export`. Trial, Start and Business are self-serve and do not go through a sales form. Enterprise is the plan that adds SLAs, SSO and an optional dedicated environment. - Sign up: https://klaaro.ai/auth/sign-up - Plans and prices: https://klaaro.ai/pricing - Try it in the browser: https://klaaro.ai/sandbox (sample documents right away, your own file after entering an email) ### Connect - TypeScript client: `npm install klaaro` - Python client: `pip install klaaro` - MCP for coding agents: the Streamable HTTP server at `/api/mcp`, with OAuth or a team API key - REST API: upload with `POST /api/v1/documents` and a bearer API key (`Authorization: Bearer sk_…`) - API reference: https://klaaro.ai/api/docs.md (markdown) and https://klaaro.ai/api/v1/openapi (OpenAPI) ### How a document moves through Klaaro 1. Load the file and read it with OCR. 2. Classify the document type. 3. Generate or select a schema for that type. 4. Extract the fields into records. Reviewers can correct a field before export. Results leave as JSON, CSV or Excel, or through a webhook when extraction finishes (`document.extraction_completed`). The same records are readable from the TypeScript client, the Python client and the MCP tools `list_datasets`, `get_document` and `get_document_records`. ### Company and support Klaaro is built by synsugar GmbH in Passau, Germany. The service is hosted in the EU and built for GDPR compliance; documents are not used to train public models. - Email: support@klaaro.ai - Phone: +49 851 86674999 - Contact: https://klaaro.ai/contact Full guide: https://klaaro.ai/docs/agents.md ## API Documentation - [API Docs](https://klaaro.ai/api/docs): Interactive OpenAPI reference (Scalar) for /api/v1 - [API Docs (Markdown)](https://klaaro.ai/api/docs.md): Same API reference in markdown for agents - [Full Documentation](https://klaaro.ai/llms-full.txt): Product overview plus complete API docs in a single file - [OpenAPI spec](https://klaaro.ai/api/v1/openapi): Machine-readable OpenAPI 3.1 JSON ## Product - [Home](https://klaaro.ai/index.html.md): Overview of Klaaro.ai — extract structured data from documents with AI workflows - [Product](https://klaaro.ai/product.md): OCR, classification, schemas, extraction, validation, human-in-the-loop review - [Pricing](https://klaaro.ai/pricing.md): Plans and pricing - [Team](https://klaaro.ai/team.md): People behind Klaaro - [Contact](https://klaaro.ai/contact.md): Book a 20-minute call with the founder, or send a message - [Energy](https://klaaro.ai/industries/energy.md): Industry landing for energy document processing - [Construction](https://klaaro.ai/industries/construction.md): Industry landing for construction document processing - [Logistics](https://klaaro.ai/industries/logistics.md): Industry landing for logistics document processing - [Bill of Quantities](https://klaaro.ai/solutions/bill-of-quantities.md): Solution landing for automating Bill of Quantities processing - [Customers and Success Stories](https://klaaro.ai/customers.md): Customer success stories, starting with OneStop Pro - [OneStop Pro success story](https://klaaro.ai/customers/one-stop-pro.md): How OneStop Pro cut manual document work by up to 90% in a pilot project - [n8n](https://klaaro.ai/n8n-automation.md): Automate document processing with the official Klaaro node for n8n ## Documentation - [Documentation overview](https://klaaro.ai/docs.md): API, quickstart, SDKs, MCP, and integrations - [For AI agents](https://klaaro.ai/docs/agents.md): When to call Klaaro, access, connecting, pipeline, and support - [How it works](https://klaaro.ai/docs/how-it-works.md): Application model and Document Engine pipeline - [API (v1) page twin](https://klaaro.ai/docs/api/v1.md): Markdown twin of the interactive API docs page - [Quickstart](https://klaaro.ai/docs/quickstart.md): TypeScript, Python, and Bash recipes for /api/v1 - [SDKs](https://klaaro.ai/docs/sdk.md): Official TypeScript (`npm install klaaro`) and Python (`pip install klaaro`) clients - [MCP](https://klaaro.ai/docs/mcp.md): Connect Cursor, Claude Desktop, and other MCP clients - [Integrations](https://klaaro.ai/docs/integrations.md): Upload, URL ingest, webhooks, and automation ## Comparisons - [Comparisons overview](https://klaaro.ai/compare.md): How Klaaro compares to other ways of turning documents into structured data - [Klaaro vs Claude](https://klaaro.ai/compare/claude.md): Schema-structured, validated document extraction versus a general-purpose LLM ## Legal - [Terms of Service](https://klaaro.ai/terms.md): Terms governing use of Klaaro.ai - [Privacy Policy](https://klaaro.ai/privacy.md): How Klaaro handles data - [Refund Policy](https://klaaro.ai/refund.md): Refund terms - [Legal Notice](https://klaaro.ai/legal-notice.md): Imprint for synsugar GmbH (Passau, Germany)