docreadi-mcp

DocReadi is document intelligence for finance teams that need structured data extracted from invoices, receipts, delivery notes, purchase orders, credit notes, and other finance documents. It turns those documents into structured rows and sends them into accounting systems and other downstream tools.

The service reads standard document types out of the box and also supports custom templates for document types not listed. It extracts every field on a document, then checks it against counterparty information it has already learned. When details drift, such as a new bank account, a new address, or totals that do not match line items, those fields are flagged for review. Approved documents can be returned as JSON through the API, sent through webhooks on state changes, exported as CSV or XLSX, or delivered as rows in an accounting system. The product also reports extraction confidence per field and keeps flagged documents in a reviewer UI where teams can approve, reject, or edit them.

Documents can be submitted through the web, the API, forwarded email, or WhatsApp after connecting a Meta Business number. DocReadi accepts PDFs, images, scans, and most modern document formats. Its API supports GET and POST requests, webhook events, idempotent endpoints, bearer-token authentication, and documented rate limits. Webhooks use signed payloads and retries with exponential backoff. Exports can be filtered by counterparty, date range, and status, and scheduled to S3 or email. Native support is listed for Xero via CSV export, while other systems can use webhooks and CSV or XLSX imports.

Pricing includes a free 90-day trial. The page also states that customer documents are not used to train shared models. DocReadi hosts tenant data in the EU, stores it in Postgres with row-level security between tenants, encrypts sensitive fields at rest, and offers two-factor authentication on every account.