Products
Each one started as a real frustration. The products first — the open source work sits below.
Live products are linked; the rest are marked as in development.
The problem: parents preparing tests or routine practice cannot find worksheets that match their child's specific needs and weaknesses, and hand-making every sheet does not scale.
The solution: a web app that generates level-appropriate worksheets, shares them through private child play links, and closes the loop with hints, scoring, and privacy controls.
The problem: AI-made browser games get built in an evening and then vanish — there is no home for them and no way for players to find them.
The solution: an uncurated public gallery where makers paste a playable URL or a small static zip, and players browse, play, favorite, and leave notes. No review queue.
The problem: password managers demand accounts, cloud sync, and subscriptions for secrets that should never leave your devices.
The solution: a local-first mobile vault with biometric access, offline password health, and an explicit Android-to-Mac companion path. No account, no cloud.
The problem: one social post is hard enough — trends shift and it is unclear what to film for TikTok, Instagram, or Shorts. Posting consistently is even harder without a marketing team.
The solution: a web app that turns one product into a ready-to-film 30-day TikTok and Instagram system — hooks, scripts, shot lists, captions, hashtags, CTAs — with weekly AI reviews that shape the next plan.
Bonsai Terminal, Andura UI, and Runpin are public today. WriteSense, Routine, and DataLens open up as they mature.
The problem: work that outlives a single HTTP request — a long AI job, a batch import — has nowhere to run, so you either hold the connection open or hand-roll a queue.
The solution: a self-hosted durable job runner. Enqueue over HTTP and get a 202 back; a worker with leases, heartbeats, and retry backoff runs it on your own machine, with cron schedules and an operator UI.
The problem: cross-platform teams duplicate design decisions four times until Flutter, React, Compose, and SwiftUI drift apart.
The solution: one shared component contract and visual catalog feeding native adapters for all four platforms.
The problem: parallel AI coding sessions dissolve into terminal chaos, tangled branches, and unreviewed changes.
The solution: a Tauri desktop control room where every session runs in an isolated Git worktree, with pinned context, coordinator and verifier flows, local history, and review.
The problem: writing coaches either upload everything you type to a cloud or give advice that ignores how you actually write.
The solution: a privacy-first macOS coach that learns from corrections locally, with optional on-device Deep Review on compatible Macs.
The problem: planners record intentions but never the reality, so you cannot learn why the plan keeps slipping.
The solution: a local-first Flutter planner pairing recurring plans and native reminders with fast event logging and early schedule insights.
The problem: data overage alerts arrive too late, and carrier apps hide which app actually consumed the quota.
The solution: an Android monitor built on OS counters — device and per-app history, plan alerts, anomaly detection, widgets, and export, all stored locally.
andura.io