Skip to content

donevaapp/android

Repository files navigation

Doneva

Automate anything on your Android phone with just your voice.

Doneva is an open-source Android app that lets you create powerful automations using natural language. Tell it what you want — "When I get a WhatsApp after 6pm, reply that I'm busy" — and it builds the automation for you. No coding, no flowcharts, no drag-and-drop. Just say it.

Download APK · Website · Privacy Policy


How It Works

  1. Speak or type what you want to automate, in plain English
  2. AI parses your request and creates a structured automation
  3. Review and confirm — Doneva shows you exactly what it will do
  4. It runs 24/7 in the background, with zero AI involved at execution time

The AI is only used once — when you create the task. After that, your automation runs as fast, deterministic code. No latency, no API costs at runtime.


What Can You Automate?

Doneva can respond to things happening on your phone and take actions automatically.

Triggers — things Doneva listens for:

  • Notifications from any app (WhatsApp, Instagram, Slack, Gmail…)
  • Incoming SMS or phone calls
  • Time of day or recurring schedules
  • Phone events (screen on/off, battery low, WiFi connected)
  • Webhooks from external services
  • Changes on a webpage
  • New emails in Gmail, calendar events, Google Sheets rows, GitHub activity

Actions — things Doneva can do:

  • Reply to notifications
  • Send SMS, make calls
  • Open apps, toggle settings (WiFi, Bluetooth, DND)
  • Send data to 27 connected services (Slack, Discord, Telegram, Notion, Trello, Google Sheets, and more)
  • Run AI-powered actions (summarise, classify, draft responses)
  • Capture and transcribe audio
  • Share content to any app

Examples:

  • "When I get a WhatsApp from my boss, reply that I'll get back in 10 minutes"
  • "Every morning at 8am, send me a weather notification"
  • "When my battery drops below 20%, turn off WiFi and Bluetooth"
  • "When I get a work email after 9pm, auto-reply that I'll respond tomorrow"
  • "When someone stars my GitHub repo, post a message in my Discord server"

AI Providers

Doneva uses cloud AI providers to parse your natural-language requests into structured automations. Bring your own API key for any of these providers:

Provider Description
Groq (recommended) Free API key, no credit card required, fastest inference
OpenAI GPT-4o Best accuracy
Anthropic Claude Great reasoning
DeepSeek Cost efficient

Your API keys are stored in hardware-backed Android Keystore on your device — they are never sent to Doneva's servers.

Voice input is powered by Whisper, running entirely on-device.


Privacy

Doneva takes your privacy seriously.

  • Notifications, messages, contacts, and call logs are processed on-device only — never transmitted to Doneva's servers
  • Voice recordings are transcribed locally using on-device Whisper and discarded immediately
  • API keys are stored in hardware-backed Android Keystore
  • Task creation queries are sent to your chosen AI provider (Groq, OpenAI, Anthropic, or DeepSeek) for parsing — Doneva does not see these requests
  • No analytics on what you automate — only anonymous app usage events (like "task created")
  • Optional cloud sync for task definitions if you create an account (you don't have to)
  • Optional Doneva Learning Network for anonymised query sharing to improve the AI (explicitly opt-in, off by default)

The full privacy policy covers DPDP (India), GDPR (EU/UK), CCPA (California), LGPD (Brazil), and PIPEDA (Canada): doneva.app/privacy


Marketplace

Browse and install community-built automations from the in-app marketplace, or publish your own.

  • 6 categories: Messaging, Productivity, Social Media, Phone Management, Daily Routines, Utilities
  • One-tap install of any community task
  • Publish your own automations for others to use
  • Browse the marketplace on the web at doneva.app/marketplace

Connectors

Doneva connects to 27 external services:

Slack · Discord · Telegram · Microsoft Teams · Notion · Trello · Todoist · Airtable · Linear · Jira · Google Sheets · Google Calendar · Google Tasks · Gmail · GitHub · Firebase · Supabase · Spotify · YouTube Music · Google Home · Home Assistant · n8n · Make · IFTTT · Pushover · Email (SMTP) · Custom Webhook

Connectors support OAuth, API key, and webhook authentication depending on the service.


Tech Stack

Component Technology
Language Kotlin
UI Jetpack Compose + Material 3
Architecture MVVM + Clean Architecture
Dependency Injection Hilt
Local Database Room (SQLite)
AI Task Parsing Cloud BYOK (Groq, OpenAI, Anthropic Claude, DeepSeek)
Speech-to-Text Whisper Base (on-device)
Networking Retrofit + OkHttp with certificate pinning
Background Processing WorkManager + Foreground Service
Auth JWT with refresh token rotation

Building from Source

Requirements:

  • Android Studio Ladybug or newer
  • JDK 17
  • Android SDK 34
  • A physical Android device running Android 10+

Steps:

  1. Clone the repository:

    git clone https://github.com/donevaapp/android.git
    cd android
  2. Copy the secrets template and fill in your values:

    cp secrets.properties.example secrets.properties
  3. Add your google-services.json from Firebase Console (project: your own Firebase project)

  4. Open the project in Android Studio and let Gradle sync

  5. Build and run:

    ./gradlew assembleDebug

Project Structure

app/src/main/java/app/doneva/
├── ai/                  # AI integration layer
│   ├── llm/             # Task parsing, prompt engineering, BYOK provider management
│   ├── action/          # AI-powered runtime actions (summarise, classify, draft)
│   ├── connector/       # 27 outbound connectors (OAuth, API key, webhook, SMTP)
│   └── web/             # Web content extraction and form interaction
├── stt/                 # On-device Whisper speech-to-text
├── analytics/           # Anonymous event tracking
├── data/                # Repositories, Room database, DataStore, API clients
│   ├── local/           # Room DB, DataStore preferences
│   ├── remote/          # Retrofit API clients, auth
│   ├── repository/      # Data access layer
│   ├── sync/            # Cloud sync for task definitions
│   └── worker/          # Background download workers (Whisper model)
├── di/                  # Hilt dependency injection modules
├── domain/              # Domain models (Task, TaskDefinition, LlmOption)
├── engine/              # Task execution engine
│   ├── trigger/         # Trigger detection and matching
│   ├── action/          # Action executors (SMS, notification, app control…)
│   ├── orchestrator/    # Execution pipeline
│   ├── scheduler/       # Cron and time-based scheduling
│   ├── service/         # Foreground services (engine, accessibility)
│   └── health/          # Engine health monitoring
├── presentation/        # UI layer (Jetpack Compose screens)
│   ├── screens/         # Home, Settings, Create Task, Marketplace…
│   ├── components/      # Shared UI components, icons
│   ├── navigation/      # Nav graph and routes
│   └── theme/           # Colors, typography, design tokens
├── push/                # Push notification handling
├── update/              # In-app update logic
├── verification/        # Account verification flows
└── util/                # Shared utilities

Supported Languages

English only. Multilingual task input is planned for a future release.


Requirements

  • Android 10 (API 29) or higher
  • 3 GB RAM minimum
  • ~25 MB app size + optional Whisper model download (~40 MB)
  • An API key from one of the supported AI providers (Groq offers free keys)

Contributing

Doneva is open source and welcomes contributions. If you'd like to contribute, please open an issue first to discuss what you'd like to change.


Licence

MIT — see LICENSE for details.

The Android app is open source. The backend API, admin dashboard, and server infrastructure are proprietary and not included in this repository.

About

Doneva — Tell your phone what to do. Open-source Android automation app with natural language task creation, 27+ connectors, AI actions, and a community marketplace.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages