Arkivra
Open source Self-hostable

Keep your documents organized and searchable.

Arkivra is an open-source document management system built for self-hosting. Organize personal records and scanned documents in structured vaults. Search their contents or enable optional AI tools when needed.

Self-host Arkivra View on GitHub
Arkivra vaults grid view

Keep your archive useful as it grows.

Arkivra gives every record a clear place and makes extracted content searchable. Optional AI adds semantic retrieval and document chat without replacing the core workflow.

Personal Vault
Home
home-insurance.pdf
lease-agreement.pdf
Family
passports.docx
emergency-contacts.txt

Organize

Use vaults to separate different areas of life. Add folders, tags and metadata for more structure.

Zahlungsfrist
service agreement.pdf
payment deadlinedue datenet 30

Search

Search extracted text instead of relying on file names. Use filters to narrow larger result sets.

What are the key terms in this contract?
AI summary

Understand

When AI is enabled, search by meaning and ask questions across individual documents or entire vaults.

Features

Tools for day-to-day document management.

Arkivra combines organization and version history with reliable retrieval through one consistent workflow.

Arkivra vault list workspace - Light Mode

Organize documents with a clear hierarchy

Vaults separate major categories. Folders provide hierarchy while tags and metadata add context. Version history preserves earlier content when a document changes.

  • Vault organization

    Separate personal and household records into dedicated vaults.

  • Tags & metadata

    Add context for filtering and group related records without changing the folder structure.

  • Version history

    Upload a replacement while keeping earlier versions available for review or restore.

  • Access control

    Control which accounts can view or update each vault.

Search document contents directly

Full-text search works without an AI provider and covers content extracted during processing. Optional AI adds semantic retrieval, document chat and PDF translation.

Core search

Available without configuring AI.

  • Full-text search

    Find exact words and phrases in text extracted during document processing.

  • Scanned documents

    Turn supported scanned PDFs and images into content you can search.

  • Filters & metadata

    Narrow results by vault, tag, metadata, or processing state to reach the right record faster.

Optional AI tools

Enabled only after an administrator configures them.

  • Semantic search

    Find passages with related meaning even when they use different wording.

  • Chat with documents

    Ask questions across selected documents, individual vaults, or all records you can access.

  • Document translation

    Translate supported PDFs between English and German using the configured model.

Use an endpoint you operate or a supported cloud provider. The selected provider determines where model requests are processed.

Read the documentation
Arkivra document chat and search workspace - Light Mode
About Arkivra

Designed for self-hosted environments.

Deploy Arkivra with Docker on infrastructure you manage. Configure vault access and storage for your deployment. Backups and optional integrations remain under operator control.

Open source

Arkivra is released under AGPL-3.0. Review the code or adapt it to your setup.

Self-hostable

Deploy with Docker on a home server or any Docker-capable host you manage.

Flexible AI setup

Use an AI endpoint you operate or connect a supported cloud provider. AI can also remain disabled.

Encrypted file storage

Uploaded originals and stored extracted asset files are encrypted at rest using keys configured for your deployment.

Activity and audit

Activity and audit logs record document actions. They also cover access, permission and administrative events.

Encrypted backups

Create encrypted backup sets containing PostgreSQL and document storage for Arkivra's restore workflow.

FAQ

Frequently asked questions

Answers about deployment, document processing and optional AI.

Does Arkivra require AI?
No. You can upload and organize files, search extracted content, manage versions, recover deleted items, and create backups without connecting an AI provider.
What does optional AI mean?
An administrator must configure a supported provider, choose the models, and enable AI before anyone can use semantic search, document chat, or PDF translation. The core document workflow remains available without it.
Where does Arkivra process documents?
You choose where Arkivra and its required Docling processing endpoint run. Core document management and full-text search do not use an AI provider. Content may leave that environment if you configure remote processing, AI, storage, or backup services.
What happens when I use a remote AI provider?
Arkivra sends the content needed for the feature you use. That can include query text and document chunks for semantic search, document context and conversation history for chat, or selected text and page images for translation. The provider may process or retain that data under its own policies.
What is semantic search?
Regular search looks for matching words. Semantic search looks for matching meaning, so a search for car insurance may also find documents that mention vehicle coverage.
Can Arkivra search scanned PDFs and documents?
Yes. Arkivra can extract searchable text from supported scanned PDFs and images. The result depends on scan quality and how the document processing pipeline is configured.
What license is Arkivra released under?
Arkivra is open source under the AGPL-3.0 license. The source code, issue tracker, and project roadmap are available on GitHub.

Ready to run Arkivra on your own server?

Use the Docker guide to prepare the required services and start Arkivra. Review the complete source on GitHub before deployment.

Self-host Arkivra View on GitHub