Why Rosetta

Most document tools are built for a quick pass: upload, translate, export, forget. Rosetta is for the documents that stay with you a little longer.

It keeps long-form PDFs and text files inside one continuous workflow: library, reader, translation pipeline, index, and source-grounded chat. Import a document once, then keep working from the same material over time.

Design Principles

• Local first. Documents, indexes, prompts, and generated artifacts live on your machine, in your system app data directory.
• Structure-aware. Rosetta tries to preserve layout and document structure before flattening everything into plain text.
• Configurable where it matters. chat, translate, embed, and rerank can each use different providers, models, and runtime settings.
• Built for return visits. A document can still be translated, indexed, searched, extracted, and discussed long after it was imported.

What You Can Do

• Build a persistent library with folders, categories, tags, and reusable metadata extraction templates.
• Parse PDFs through MinerU, using the built-in runtime, a self-hosted service, or the official MinerU API.
• Read original and translated content in separate or side-by-side views.
• Route chat, translate, embed, and rerank work through dedicated provider channels.
• Tune translation prompts, chunking strategy, concurrency, rate limits, and failover behavior.
• Search your library with both direct matching and semantic retrieval.
• Ask questions against documents and inspect the supporting sources.
• Track parse, translation, and indexing jobs in a unified task center.
• Export local backups or sync workspace data over WebDAV.

Good Fit

• Research papers, standards, manuals, and technical reports you revisit over weeks or months
• Internal documentation that benefits from bilingual reading and reliable source tracking
• Personal or internal knowledge bases built from real documents instead of loose notes
• Translation workflows where terminology, provenance, and later retrieval all matter

Quick Start

1. Download

Get the latest release from GitHub Releases.

2. First Run

1. Install and open Rosetta.
2. Go to Settings and configure at least one AI provider channel.
3. Choose a MinerU mode: built-in, external, or official API.
4. Import a document into your library.
5. Run Parse -> Translate -> Index.
6. Continue in Search, Chat, or the document reader.

3. Typical Workflow

Import -> Parse -> Translate -> Index -> Read / Search / Chat -> Backup

Build from Source

Requirements:

• Node.js 18+
• Rust stable
• Tauri 2 prerequisites
• Python 3.x for the built-in MinerU runtime or the optional zvec backend

git clone https://github.com/somnifex/Rosetta.git
cd Rosetta
npm install
npm run tauri:dev

Data and Privacy

Rosetta is local-first by default. Runtime data is stored in your system app data directory rather than inside the repository workspace.

Typical stored data includes:

• imported document copies
• SQLite metadata and indexes
• MinerU runtime files and downloaded models
• local cache, generated outputs, and backup artifacts

Rosetta only talks to the external services you configure. You decide which providers to use and what content leaves the machine.

Roadmap

• Document library management
• MinerU integration
• Multi-channel model routing
• Search and source-grounded chat
• WebDAV sync and local backup
• Multilingual interface
• Batch translation workflows
• Richer annotation and review tools
• More export and collaboration options

Contributing

Contributions are welcome in UX, reliability, parsing quality, translation quality, documentation, and platform support.

Before opening a pull request, please run:

npm run build
cd src-tauri && cargo check

License

Rosetta is licensed under the GNU General Public License v3.0.

Acknowledgments

• Tauri
• MinerU
• Radix UI
• shadcn/ui
• Rosetta also owes a quiet debt to the issue threads, forum posts, and troubleshooting notes shared across GitHub, Reddit, and Linux.do.