Your prompts, curated and your own, usable in any app.
Loadstash is a local-first Android app that keeps your best AI prompts one tap away — inside any app on your phone. Select text, tap Loadstash in the long-press menu, pick a prompt, fill in variables, and it drops straight into the field.
- Works in any app — integrates via Android's
ACTION_PROCESS_TEXT(long-press selection menu) and a keyboard-triggered floating bubble - Variable fill-in — wrap any part of a prompt in
{{double braces}}and Loadstash shows a fill-in sheet before inserting - Folder organisation — nested folder structure (Writing › Email › Cold), breadcrumb navigation
- App-aware ranking — prompts used in Claude surface first in Claude; prompts used in ChatGPT surface first in ChatGPT
- Custom model tags — tag prompts for Claude, ChatGPT, Gemini, or any model you add, with custom colours
- Quick Add — save a prompt directly from the overlay without switching to the main app
- APM-compatible import/export — share prompt packs as
.zipfiles using the Agent Package Manager.prompt.mdformat - Light & dark theme
- 100% local — nothing leaves your device
- Flutter 3.x (
flutter --version) - Android device or emulator running API 26+
- Android Studio or VS Code with Flutter extension
git clone https://github.com/your-username/loadstash.git
cd loadstash
flutter pub get
dart run build_runner build --delete-conflicting-outputs
flutter runflutter testlib/
core/ — theme tokens, animation primitives
data/ — Drift SQLite database, repositories
features/ — screens (library, overlay, editor, onboarding, settings)
providers/ — Riverpod state providers
services/ — business logic (variable detection, import/export, model tags)
android/ — native Android (BubbleService, AccessibilityService, ProcessTextActivity)
Key technologies: Flutter · Dart · Drift (SQLite) · Riverpod · go_router
We welcome contributions of all kinds — bug fixes, new features, documentation, and starter prompt packs. See CONTRIBUTING.md.
For security vulnerabilities, see SECURITY.md and follow the responsible disclosure process.
Copyright 2026 Kushal Agrawal. Licensed under the Apache License, Version 2.0.