L//IOS: An Operating System for Insight.
Three applications built on one idea: software that understands the data, says what matters, and shows its working.
Less navigating. More understanding.
Modern software asks people to move between fragmented systems, interpret information by hand, and repeat the same analytical steps again and again.
L//IOS explores a different approach: applications that do more of the interpreting, keep context across questions, and help people act with less friction, without ever inventing a number.
The future of analytics isn’t more dashboards. It’s systems that understand the data and tell us what matters.
Each app runs the same five steps on a different kind of data.
Three applications, one way of thinking.
These are real screenshots, captured by running each app’s own QA on demo or synthetic data. None of the data shown is real.
L//IOS Analyst
Drop data → Initialize → Understand → Dashboard → Insights → Investigate
Give it files it has never seen. It works out the structure, connects the datasets safely, checks quality, builds the dashboard and investigates what changed. Every number traces back to source rows.
- Excel, CSV, JSON, Parquet and SQLite ingestion, including messy sheetsWorking
- Schema and meaning detection: measures, dimensions, dates, hierarchiesWorking
- Relationship discovery with match rates; joins can never multiply rowsWorking
- Data quality checks with stated methods and affected rowsWorking
- Generated dashboard, driver trees and ranked findingsWorking
- Plain-language questions answered by deterministic toolsWorking
- Version refresh: what changed between data dropsWorking
- PostgreSQL / SQL Server snapshotsPrototype
- Optional AI rephrasing and summaries, numerically guardedPrototype
- 166,286 synthetic trades, 5 files, 9 data regions understood in under 25 seconds
- Engine totals checked against an independent recomputation from the raw workbook
- Measured on a 72 MB workbook: 831,430 rows ready to explore in about 87 s
- Raw data never leaves the computer; AI is off by default
The rules every L//IOS app follows.
- 01
Code calculates. Models interpret.
Arithmetic is deterministic and tested. A model may explain, but every number it writes must already be in the evidence.
- 02
Show the working.
Every KPI, bar and finding opens Explain, Drill down and View data, back to the source row.
- 03
Check the data first.
Before a change is called a business event, the data explanations are tested: missing loads, duplicates, mapping changes.
- 04
Private by default.
Raw data stays on the device. AI is off until you turn it on, and you can see exactly what it would be sent.
The foundations.
Technologies in the current builds. Not every application uses every one.
| Technology | Analyst | Markets | Health |
|---|---|---|---|
| TypeScript | used | used | not used |
| React | used | used | not used |
| Node.js | used | used | not used |
| SQLite | used | used | used |
| esbuild | used | used | not used |
| Kotlin | not used | not used | used |
| Jetpack Compose | not used | not used | used |
| Health Connect | not used | not used | used |
| Anthropic API | not used | used | not used |
| Playwright | used | used | not used |
Where it’s heading.
These are directions, not finished capabilities.
- NextDesktop and phone, one context
Link L//IOS Mobile to desktop L//IOS so the brief follows you.
- NextMore data sources
Fundamentals, more attention sources, and live database connections for Analyst.
- NextResearch that stays current
Scheduled revisits of saved research: “What changed today?”
- NextGuarded AI, everywhere
Models that plan and explain, never calculate, with every number checked against evidence.
- NextKnowledge graph
Questions across entities, relationships and themes the research has already found.
- NextNutrition, complete
Finish nutrition in L//IOS Health and connect it to training and recovery.
L//IOS is an independent personal technology project and is not affiliated with or endorsed by RBC Capital Markets.