Kalandar

Kalandar is a natural-language capture parser for tasks, reminders, calendar events, and low-signal thoughts. It is built around one capture-preview contract with JavaScript, Rust, and Swift/Core AI implementation paths.

Interactive demo

Try the capture parser

This demo calls parseCapturePreviewInput from Kalandar on the server and turns plain language into a structured preview before anything gets saved.

Parsing capture text.

Kind
Parsing
Candidates
Parsing
Warnings
Parsing
Status
Released product
Date
July 2026
What it is
A parser and runtime contract that turns messy capture text into a structured preview before anything becomes a task, reminder, calendar event, or thought.
Why it was important
It makes the first step of productivity software inspectable. You can write what you mean in plain language, review the structured result, and only then commit it into a real system.
Architecture
One capture-preview contract sits behind JavaScript, Rust and Swift/Core AI implementation paths, so the same plain-language input can be inspected in web, cross-platform native and Apple-native contexts.
Access angle
Capture should not require memorising rigid syntax before the idea is even saved. Kalandar treats plain language as the input surface and structured preview as the accessibility layer.
Role
Product design, architecture, implementation, documentation, release and demo design.
Tools used
TypeScript, Rust, Swift, Core AI, Codex and Claude Webapp.
Links
GitHub repository and v0.1.2 release

Back to projects

Blinding Pixels · GitHub