Docspell

Docspell is a self-hosted, open-source document management server purpose-built for organising piles of paper that have already been scanned, downloaded, or received by e-mail. Where commercial document SaaS services trade convenience for visibility into the most sensitive corners of personal and small-business life — tax filings, medical records, legal correspondence, contracts, receipts — Docspell keeps the entire archive on infrastructure you control while still delivering OCR, NLP extraction, full-text search, and rich tagging out of the box. The result is a private, queryable filing cabinet that scales from a single user's household paperwork to multi-tenant workspaces shared across small teams.

Common Use Cases

Households use Docspell as a centralised home for tax forms, utility bills, insurance documents, school correspondence, and warranty paperwork — letting any family member upload via the web UI, e-mail attachments to a configured address, or drop scans into a synced folder. Freelancers and consultants run Docspell as the canonical archive for invoices, contracts, and expense receipts, taking advantage of NLP-extracted amounts and dates to track project profitability without manual data entry. Small businesses self-host Docspell for AP/AR record-keeping, audit-trail-friendly invoice storage, and shared access across an accounting team while keeping vendor data inside the company's trust boundary. Genealogists and family archivists scan multi-decade collections of letters, certificates, and photos into Docspell so OCR makes handwritten records discoverable and the metadata extraction surfaces dates and named entities. Privacy-conscious users self-host Docspell to handle medical records, mental-health correspondence, fertility data, and legal documents that they specifically do not want flowing through a SaaS document graph. Small clinics, law offices, and consulting firms deploy Docspell as a multi-collective instance — separate workspaces per client or matter — without the per-seat fees commercial document platforms impose.

Key Features

• OCR pipeline using Tesseract that processes scanned PDFs, images, and faxes into searchable text on ingest
• NLP extraction of dates, correspondents, amounts, and contacts that pre-classifies documents in the inbox
• Solr-backed full-text search with filters by tag, folder, custom field, date range, and correspondent
• Mail-fetch integration polling IMAP mailboxes for attachments to import automatically
• Integration HTTP endpoint for scanner companion apps and custom upload pipelines
• Document conversion via WeasyPrint, Unoconv, and Tesseract turning HTML, Office, email, and image files into searchable PDFs
• Multi-collective architecture giving each workspace isolated users, documents, tags, and metadata
• Tagging system with categorical tags, custom fields, organisations, and persons for rich annotation
• Saved queries that auto-refresh as new matching documents arrive in the archive
• Folder hierarchy plus drag-and-drop reorganisation directly inside the web UI
• Encryption-at-rest support for sensitive document collections via PostgreSQL backend configuration
• Signup-mode controls (`open`, `invite`, `closed`) so new instances can be locked down after the initial admin signs up
• REST API and CLI (`dsc` tool) for automation, bulk imports, and integration with external systems

Why deploy Docspell on Hostinger VPS

Running Docspell on a Hostinger VPS gives households, freelancers, and small businesses a private document archive that does not flow through any SaaS provider. Document metadata reveals more than most users realise — financial relationships, medical history, family composition, professional engagements — and self-hosting on infrastructure you control keeps that data inside your trust boundary by default. Dedicated CPU on a VPS handles the JVM-based rest server, the joex worker that runs OCR and document conversion, and the Solr index simultaneously without throttling, even during bulk imports. Persistent volume storage keeps the PostgreSQL database, Solr full-text index, and uploaded documents durable across container restarts and host upgrades — important for a system whose value compounds as years of records accumulate. Combined with Traefik-based HTTPS routing on a clean branded hostname, the upload UI and IMAP-fetch traffic stay TLS-encrypted, and the integration endpoint that scanners and CLI tools post to is also reachable over HTTPS. The four-service footprint (rest server, joex, Postgres, Solr) is heavier than a single-binary tool, so a mid-tier VPS plan is a better fit than the smallest plans, but the trade-off buys a meaningfully more capable archive than lighter document tools can provide.