Show HN: VelocityNote – A tiny Markdown notebook with local AI
VelocityNote introduces an AI agent-driven import system that converts notes from multiple formats (Notion, Word, PDF, Evernote) into a unified Markdown folder structure The agent performs a dry-run first, showing a preview of what will be imported (note count, images, notebooks) before any changes are made Re-importing the same content updates existing notes rather than creating duplicates, using a content digest for change detection Original metadata such as dates, images, and folder hierarchy
Analysis
TL;DR
- VelocityNote introduces an AI agent-driven import system that converts notes from multiple formats (Notion, Word, PDF, Evernote) into a unified Markdown folder structure
- The agent performs a dry-run first, showing a preview of what will be imported (note count, images, notebooks) before any changes are made
- Re-importing the same content updates existing notes rather than creating duplicates, using a content digest for change detection
- Original metadata such as dates, images, and folder hierarchy are preserved during conversion
- The system handles mixed-source imports in a single pass, consolidating disparate note formats into one native Markdown workflow
Why It Matters
This represents a practical application of AI agents in personal knowledge management, addressing a common pain point: the fragmentation of notes across multiple platforms and formats. For AI practitioners, it demonstrates a clean agent workflow pattern—dry run, review, confirm, execute—that could be adapted for other data migration or transformation tasks.
Technical Details
- The agent accepts natural language commands to import from mixed sources (e.g., Word, PDF, Evernote .enex files) into a staging directory before final import
- A dry-run mode (
import_folder(dry_run: true)) previews the import without making changes, reporting counts of notes, images, and notebooks - Content digest-based deduplication ensures re-imports update existing notes rather than duplicating them
- The system skips non-note files (e.g., .xlsx spreadsheets) with warnings, showing selective processing logic
- Supports Obsidian vault imports with wiki-links and attachments, maintaining folder structure natively
Industry Insight
- The "dry run → review → confirm" agent pattern is a robust template for building trustworthy AI tools that handle user data, reducing the risk of accidental data loss or corruption
- Unified import pipelines that normalize diverse formats into a single standard (Markdown) could become a key differentiator for note-taking and knowledge management platforms
- The emphasis on metadata preservation (dates, images, folder hierarchy) highlights that users prioritize fidelity over convenience when migrating content—tools that ignore this risk losing adoption
Disclaimer: The above content is generated by AI and is for reference only.