All selected work

Developer tooling · Local-first visibility

Codex Usage

Local developer-tooling visibility without a cloud dashboard.

A Quickshell module that combines local Codex session totals with live, cached, stale, or local-only rate-limit state in one compact interface.

Public repository · Documented state model
Developer tooling · QML + Python
Technologies
QML · Quickshell · Python · Local data processing
State model derived from the public implementation and documentation; no private usage data is shown.

01

The state problem

Useful usage data arrives from different local sources and at different freshness levels. A single number without source state can look more authoritative than it is.

02

System shape

  1. Local session metadata
  2. Optional live limit snapshot
  3. Collector
  4. Cache and freshness state
  5. QML service
  6. Bar indicator and detail panel

03

Key engineering decisions

01

Expose freshness

The interface distinguishes live, cached, stale, and local-only states.

02

Degrade gracefully

Cached and local session metadata remain useful when the live source is unavailable.

03

Keep workspace collection opt-in

Workspace names are not collected by default.

04

Respect disabled state

The collector pauses when the module is disabled.

05

Follow the desktop environment

Theme-aware UI and configurable palettes integrate with the existing Quickshell surface.

04

Result and evidence

A publicly inspectable, installable module with explicit privacy controls, source-state feedback, and documented fallback behavior.

05

Scope boundaries

  • This is an independent module, not an official OpenAI integration.
  • Real token totals, reset times, workspace names, account identifiers, and local paths are not shown here.
  • The visual on this page is a documented state model, not a product screenshot.

Available for the right team

Open to remote product engineering roles

I am looking for a stable remote role where I can own meaningful product or engineering outcomes across AI-assisted systems, Python, automation, integrations, Telegram, and web products.