#
Xentropy supports multiple capture modes. Each feeds into the same AI pipeline — transcribe, analyze, embed, connect — so no matter how you capture, the result is structured, searchable knowledge.
#
Voice is the primary capture mode. Designed for the "thinking professional" — founders, researchers, writers — who has more thoughts than time to organize them.
Key features:
- One-tap record — open the app, tap the mic, speak
- Stream-of-consciousness — ramble freely; the AI segments your monologue into atomic thoughts
- Background processing — transcription (Whisper), summarization, tagging all happen asynchronously
- Multi-platform — mobile app, lock-screen widget, share sheet
See the Voice Setup guide for platform-specific instructions.
#
The classic capture mode. Available on both web and mobile.
- Free-form text with markdown support
- Auto-saved as you type
- AI enrichment happens after save (summary, tags, folder suggestion)
#
Capture whiteboards, documents, or anything visual.
- Take a photo with the in-app camera
- Upload from your gallery
- Stored alongside text and voice notes in your Wall
Optical character recognition (OCR) for text extraction from images is on the roadmap.
#
Every capture flows through the same pipeline:
flowchart LR
Capture --> Transcribe[Whisper Transcription]
Transcribe --> Analyze[AI Analysis]
Analyze --> Embed[Vector Embedding]
Embed --> Graph[Knowledge Graph]
Graph --> Wall[Your Wall]
- Transcribe — voice is converted to text via OpenAI Whisper
- Analyze — the content gets a summary, tags, sentiment, and entity extraction
- Embed — a vector embedding is created using
text-embedding-3-small - Connect — the capture is linked to related notes via pgvector similarity search
- Organize — a folder is suggested (AI suggests, you confirm — nothing is filed without your approval)
All processing happens in the background. Your Wall updates in real-time as each stage completes.
#
| Plan | Voice Recording | Storage | |------|----------------|---------| | Free | Up to 5 min per capture | 50 MB total | | Pro | Up to 30 min per capture | 10 GB total |
Note: Pro plan limits are not yet enforced — we're currently in closed beta.