Taggard is built on a simple principle: the information you capture should never be used against you, shared without your knowledge, or held hostage. Here is exactly how we protect it - and the roadmap for giving you even more control.
Your documents are processed by AI to extract meaning, then stored encrypted. Row-Level Security means the database itself enforces that only your account can reach your data - not our team, not other users. We could not read your notes if we tried.
AES-256 encryption at rest
TLS 1.3 in transit
Row-Level Security - enforced by the database, not the application
Full data export at any time (JSON / CSV)
Deletion on request - removed within 48 hours
We do not sell data, train models on your content, or share documents with third parties.
Taggard will generate embeddings - the mathematical representation of a document's meaning - directly inside your browser using a local model. The raw text never touches our servers; only the embedding and its tags sync. We never see the source.
Powered by transformers.js - runs entirely in your browser, no plugins.
For legal, medical and government work, we are building a native desktop application. Every step runs on your machine: local models, local vector database, local storage. The internet is optional - built for professionals who cannot risk any cloud exposure.
Local AI processing - no API calls
File-system watcher auto-indexes your folders
SQLite + FAISS on-disk vector search
Zero network required
Encrypted local database
Backup is just copying a folder
Planned for Windows, macOS and Linux.
We will never sell your data.
We will never train AI models on your documents.
We will never share your content with third parties.
We will never lock you in - export your full knowledge base at any time.