Import representative material
Bring in documents, datasets, media, collections, and structured exports. MyDataSights starts with practical local viewers and analyzers, with more formats added over time.
Preview build · private local workspace
A private, offline-first,
portable knowledge workspace.
Bring in your data. Analyze locally. Find what matters.
Make sense of your own information, privately: documents, datasets, media, collections, and structured exports. Search, view, organize, analyze, report, export, share, and back up your work, locally by default.
No account for local workspace use. No product telemetry. Optional local models run on-device after setup.
- Linux desktop
- ·
- Windows desktop
- ·
- Web/self-hosted
- ·
- Responsive mobile web
- ·
- Android when stable
amoderndev.mydatasights.app
Sources
CSV air-quality.csv
CSV electrical-3phase.csv
CSV books-library.csv
MD study-guide.md
PDF field-report.pdf
JSON archive-index.json
Workspace
Search filenames, text, tags, metadata, and extracted fields.
Open files as readable documents, tables, charts, and time-series views.
Run summaries, domain insights, correlations, anomaly checks, and reports.
Open each generated result beside the source it came from.
Outputs
Review generated JSON and Markdown results as tables and documents.
Export reports, summaries, correlations, and data views.
Share useful result text or table data when needed.
Back up the workspace as a portable file.
Static preview of the current workflow. Use the live example for the real app surface.
- Local-first workspace. Keep files, extracted structure, and analysis close to the source.
- Portable backups. Export and restore `.mdsproj` workspaces instead of locking work into one install.
- No product telemetry. The app does not add analytics, tracking pixels, or product usage reporting.
- Share on your terms. Copy, export, share, or back up selected results when you choose.
From scattered files to a usable workspace
MyDataSights keeps the loop close to the source: import material, inspect what was extracted, organize the useful parts, analyze structure, create reports, export, and share only what you choose.
Search and inspect
Find content by filename, text, metadata, tags, and extracted fields. Meaning-based ranking is described only when the active model path is confirmed.
View and analyze
Use formatted data views, tables, table summaries, charts, time-series views, analyzer outputs, and reports where the current build supports them.
Connect relationships
Surface relationships between files, entities, columns, tags, time periods, sources, documents, and outputs when the workspace has enough structure.
Report, copy, export, share
Turn generated results into practical outputs: reports, exported data views, copied result text or table data, share actions, and portable workspace backups.
Local analysis and intelligence
Built-in analyzers can summarize and extract structure now; small local models can help where enabled. Larger local AI features are planned later.
Preview status
Current, preview, and planned capabilities are separated so the product can be useful now without overstating what is still being built.
| Capability | Status | What it means |
|---|---|---|
| Local workspace storage | Available | Workspaces are local by default and can be backed up as `.mdsproj` files. |
| Linux, Windows, web/self-hosted, responsive mobile web | Primary targets | These are the first stable targets for the 1.0 path. |
| Native Android app | Candidate | Can join the primary group when stable and behavior-compatible. |
| iOS and macOS apps | Planned later | Planned after the first stable target group. |
| Charts, table views, time-series, reports | Available where implemented | Current app surfaces focus on practical table, chart, time-series, and report views. |
| Semantic search and larger local AI features | Planned or experimental | Broader chat, transcription, and model-backed ranking paths remain planned or experimental. |
| Image OCR | Partial / planned | Image text extraction is target-specific today; native desktop OCR is planned before it is presented as broadly ready. |
| Third-party plugin runtime and raw-data rules | Future direction | Parsers, analyzers, themes, engines, and model metadata exist; stable external loading comes later. |
Useful across messy personal, professional, and research material
Collections
Books, board games, hobby archives, personal libraries, and collection catalogs.
Data
Sensor logs, medical exports, engineering records, time-based files, and other structured records.
Documents
Official documents, study material, reports, records, private research archives, and reference material.
Publish or explore a selected workspace on the web
MyDataSights can be deployed as a static, self-hosted workspace seeded from files you choose. Visitors can read, search, inspect, chart, and explore the published workspace.
Live example
amoderndev.mydatasights.app is the current public portal and live example of the web app behavior.
Self-hosting shape
Seed selected files, build the web workspace, and deploy it on hosting you control. The technical guide can cover the current seed directory and deployment options after the repository opens.
Visitor data
Where upload or paste is enabled, visitor data stays client-side and is not stored on the host server.
Built for private work, not hidden data flows
Desktop app. Local workspace by default. Filesystem access is controlled by the device and app permissions.
Web and live examples. Browser-based workspace behavior is separate from desktop; hosted pages may have normal provider request logs.
Model downloads. Optional model downloads fetch model files to your device. That action does not upload your workspace data.
AI and analyzers. Output is advisory and should be checked against the source, especially for medical, legal, financial, or safety-critical decisions.
Frequently asked
What is MyDataSights?
MyDataSights is a private, offline-first knowledge workspace for turning scattered information into searchable views, visible relationships, source-backed findings, reports, exports, and portable backups.
What can I try today?
The public preview points to a live self-hosted example at amoderndev.mydatasights.app. Public source and downloadable installers open with the public release.
Does anything leave my computer?
The app does not upload your workspace data for product telemetry. Optional model downloads fetch model files to your device when you choose them. Hosted web pages may have normal infrastructure logs from the hosting provider.
Do you collect analytics?
This landing site does not include product analytics, cookies, tracking pixels, or hosted error reporting. Hosting providers may still process normal request logs for delivery, security, and abuse prevention.
What file types does it support?
MyDataSights focuses on representative types of information: documents, datasets, media, collections, and structured exports. Current file examples include Markdown, PDF, HTML, EPUB, plain text, CSV, JSON, common image/audio/video files, and code. More formats, raw-data viewers, and interpretation rules are future directions.
Are AI features ready?
AI is not the lead promise of the preview. Current small-model or stable local analysis paths can be shown where they work; Phi-3 chat, Whisper transcription, and broader semantic ranking are planned or experimental until confirmed.
Which platforms are targeted?
Primary targets are Linux desktop, Windows desktop, web/self-hosted portal, and responsive mobile web. Native Android may join that group when stable and behavior-compatible. iOS and macOS are planned later.
Is it open source?
MyDataSights is intended for Apache-2.0 release. During preview the repository remains private; public source, notices, and release artifacts open with the public release.
How do I report a security issue?
Email [email protected]. Please avoid posting vulnerability details in public issue trackers.
Explore the live workspace, then follow the release.
Preview now, public source and downloads at release.