ExphoraDB · Nob

##See it work

##Fast and local

It opens JSON, JSON Lines, CSV, XML and SQLite in under two seconds. The table is virtualized: it only draws the rows on screen, so it handles millions. Double-click a cell to edit it, with autosave and undo/redo

##Calculated columns

It has its own expression engine written in Rust (a Pratt parser, expr.rs): comparisons, if(price > 100, "Premium", "Regular"), round, plus sums, averages and counts that only take the filtered rows into account

##.exh views

An .exh file stores the whole session: filters, columns, sorting, calculated columns, charts, notes and the path to the dataset. If the original file was moved, it finds it again by its SHA-256

##Sharing without servers

Datasets can be sent to another PC on the local network over end-to-end encrypted P2P, without going through any server. And exported to CSV, JSON, NDJSON, Markdown, PDF or Excel

~/exphoradb · how it works

data.csv · .json · .xml · .sqlite
 │ Rust + Tauri v2
 ▼
virtualized table (millions of rows)
 │ filters · regex · calculated columns
 ├─► .exh view (filters, charts, notes)
 ├─► export: CSV · JSON · PDF · Excel
 └─► encrypted P2P ─► another PC on the network

ExphoraDB on Windows with vulnerabilities.json open: columns cvss, discovered, id, port, service, severity, status and target — A test vulnerabilities JSON: 8 rows, 8 columns and the button to add a calculated column · development build, Mar 2026

ExphoraDB with tech_stack_survey open: 15 rows and 7 columns of programming languages, with average salary, category and year created — A programming-languages survey in another tab, with recent views on the left and "Save view" to keep it as an .exh

##Log

Mar 10, 2026First commit, already at v0.6.2The repo starts with the app already working. Same day: inline editing with autosave and undo/redo, the project page, and .exh files registered on Windows
Mar 11, 2026v0.7.1 and v0.7.2Two releases in one day
Mar 14, 2026v0.8.4The current version, plus the MIT license and a security policy