Skip to main content

Frequently asked questions

What is Jiandu?

Jiandu is a self-hosted document management system for a household or another small, trusted group. Its browser workflow adds, organizes, finds, and downloads preserved originals, while the native API manages exceptional document restrictions. The application, database, and original files remain under operator control, and a separate operator CLI creates portable exports with versioned metadata.

Is Jiandu open source?

Not yet in the legal sense. No license has been selected, so default copyright applies and the code cannot be assumed reusable or redistributable. Repository visibility does not change that legal boundary. If reuse or redistribution matters to your decision, check for a LICENSE file first.

Is Jiandu production-ready?

Jiandu does not make a blanket production-readiness claim. The repository is evolving, release packages are not currently published, and some release evidence is incomplete. Evaluate with non-critical samples, review the support matrix, and prove backup and restore on your exact deployment.

Does Jiandu require a cloud account?

No hosted Jiandu account is required. The default profile uses local disk and SQLite and sends no remote telemetry. You can opt into systems you operate or choose, including S3-compatible storage, WebDAV, OIDC, LDAP/AD, email connectors, and OTLP telemetry.

Where are documents stored?

Original and durable document blobs can live on local disk, S3 or S3-compatible object storage, or WebDAV services including ownCloud and Nextcloud. Searchable records live in SQLite by default or PostgreSQL 17.x or 18.x. A complete recovery plan protects both the database and blobs.

Does Jiandu support OCR?

A package may include an optional local OCR worker on supported platforms, but Jiandu does not currently advertise an end-to-end OCR accuracy, latency, memory, or failure-rate guarantee. Remote OCR is not supported by the household profile. Preserve originals and expect to correct searchable metadata for image-only documents.

Is Jiandu a Paperless-ngx replacement?

Jiandu does not claim drop-in replacement compatibility. It provides a reviewed import path for supported Paperless exports, with explicit preflight losses and a verification workflow. Keep the Paperless source available and read the migration guide before importing.

Can more than one Jiandu server use the same installation?

No. The supported topology is one Jiandu application process per installation. Do not run active-active replicas or multiple workers against the same database and blob namespace.

Can I access Jiandu from another device?

Yes, through an explicitly configured HTTPS reverse proxy in front of Jiandu's private HTTP listener. Do not publish the default plaintext listener. Follow network access and HTTPS and verify the route from an ordinary household device.

How do household permissions work?

The library is shared by default. Owners and Administrators can use roles, groups, custom permission sets, and explicit document grants. A restricted document can grant Viewer, Editor, or Manager access to named people or groups. Administrator status alone does not reveal every restricted document; the Owner retains a recovery grant.

How do I back up Jiandu?

Use jiandu-portability backup for the supported recovery archive, then run verify-backup after copying it. The archive contains the database and referenced blobs but deliberately excludes configuration, secret files, and the Owner token; protect those separately through their appropriate secret and recovery channels. A backup is unproven until it has been restored into a new isolated destination and representative originals open. Follow backup and restore for the exact boundaries.

What should I try first?

Use the quickstart, add one non-sensitive PDF or image, correct its title, find it without using the filename, and download the preserved original. Before relying on the installation, follow the operator guidance to create a portable export and perform a restore drill. That small proof is more useful than importing a large archive immediately.