Turn supported document files into concise, reusable knowledge.
- Concise AI-generated summaries for supported document types
- Structured key insights you can review later
- Automatic tags and category assignment
- Brief, Standard, Detailed, and Custom analysis modes
- Useful for reports, manuals, notes, guides, and research material
Turn recordings into transcripts, summaries, and searchable notes.
- Local transcription using Whisper models
- Multiple Whisper size options for speed versus accuracy
- Brief, Standard, Detailed, and Custom post-transcription analysis
- Batch processing for multiple recordings
- Export transcript-only, summary-only, or combined output
Extract spoken content from video and turn it into something easier to navigate.
- Video audio extraction through bundled or configured FFmpeg
- Local speech-to-text transcription for spoken video content
- AI summaries and structured insights from the transcript
- Optional per-file vision assist for silent clips or videos where visual context matters
Bring image files into the same private workspace as documents and media.
- Dedicated Image Insights workflow in the desktop app
- Local vision-model-backed image analysis without cloud upload
- Selectable image models, currently including Moondream2 and LLaVA 1.6 Mistral 7B
- Review, filter, search, export, and manage image results alongside other insight types
- AI rename support for analyzed images
Synthesize related analyzed files into a saved higher-level summary.
- Select at least two analyzed files and create one synthesis from their existing summaries
- Successful runs are saved automatically to the Combined Insights page
- Reopen, export, and regenerate saved combined-insight runs later
- Useful for grouped review of related recordings, documents, videos, or images
Build a searchable library you can revisit and expand over time.
- Search across summaries, transcripts, insights, tags, and related metadata
- Real-time debounced search as you type
- Relevance-ranked results across media, documents, and image insights
- Jump directly from results into the relevant processed file details
Ask questions and get answers grounded in your processed files.
- Ask natural-language questions against a selected processed file
- Current chat scope is one processed file at a time
- Answers are built from stored summaries, transcript excerpts, and related content
- Conversation history persistence with brief or standard reply styles
Discover content in bulk and add files to focused insight collections.
- Recursive folder scanning through Auto Scan
- File-type filters for video, audio, documents, and images during scan
- Hierarchical file browsing with search, multi-select, pagination, and sorting
- Batch add-to-insights flow for selected files
Handle larger libraries with queue tracking and clear progress.
- Queue-based batch analysis for larger runs
- Progress indicators, status messages, and estimated time remaining
- Stop or cancel support during processing
- Retry-friendly workflows for pending or failed items
Rename files based on their analyzed content.
- AI-generated rename suggestions based on processed content
- Bulk rename workflow through a dedicated rename modal
- Review-before-apply behavior so users remain in control
- Available across analyzed documents, audio, video, and images
Export results in formats that are easy to share, save, or archive.
- Export to TXT or PDF
- Choose summary-only, transcript-only, or combined output
- Export saved combined insights as separate deliverables
- Useful for sharing, archiving, or republishing processed results elsewhere
See what is ready, what is still running, and how the app is performing.
- Dashboard totals with breakdown by video, audio, document, and image counts
- Success-rate and average-processing-time indicators
- Recent insights and processing history with refresh, paging, selective delete, and clear-all actions
- AI, Whisper, image-model, and GPU readiness indicators
Choose models, hardware options, and app behavior to match your setup.
- Download and manage text-generation, Whisper, and image models
- Current text-model catalog includes Gemma, Qwen, Phi, Llama, and Mistral choices at multiple size tiers
- Current image-model choices include Moondream2 and LLaVA 1.6 Mistral 7B
- Enable GPU acceleration, configure FFmpeg, and control onboarding and runtime setup
