Deutsch | Español | français | 日本語 | 한국어 | Português | Русский | 中文

The most secure way to understand your finances.

Whisper Money is a privacy-first personal finance application that helps you track, categorize, and understand your spending. We don't sell your data and we don't profile you for ads. The entire codebase is public, so you can check exactly where your data goes.

💬 Join our Community: Whether you're a user looking for help or a developer wanting to contribute, we'd love to have you in our Discord server! Share feedback, ask questions, discuss new features, or just hang out with fellow privacy enthusiasts.

• 🔐 Privacy-first — You own your data and we never sell it. Self-host it and point the AI at a local model to keep it entirely on your own infrastructure
• 🏦 Bank account management — Track multiple accounts in one place, by hand, from an import, or synced straight from your bank with open banking
• 📊 Transaction categorization — Automatic and manual categorization
• 🤖 Automation rules — Set up rules to auto-categorize transactions
• 📈 Financial insights — Understand your spending patterns

• Backend: Laravel 12, PHP 8.4
• Frontend: React 19, Inertia.js v3, TypeScript
• Styling: Tailwind CSS v4
• Database: MySQL
• Cache/Queue: Redis
• Testing: Pest v4

The easiest way to get started is using our automated setup script:

bash <(curl -fsSL https://whisper.money/setup.sh)

After installation, just visit https://whisper.money.localhost in your browser.

If you prefer to set up manually:

1. Clone the repository:

git clone https://github.com/whisper-money/whisper-money.git
cd whisper-money

1. Run the setup script:

whispermoney install

Important: You must run whispermoney install before using any other command. If you skip the install step, commands like start will not work.

Once installed, you can use the whispermoney command for common tasks:

# Start all services
whispermoney start

# Stop all services
whispermoney stop

# Upgrade to latest version
whispermoney upgrade

# Interactive menu
whispermoney

For active development with hot reloading:

composer run dev

This will concurrently start:

• PHP development server (via Portless HTTPS proxy)
• Queue worker
• Log viewer (Pail)
• Vite dev server

The application will be available at https://dev.whisper.money.localhost. In git worktrees, the branch name is automatically prepended (e.g. https://fix-ui.dev.whisper.money.localhost).

For testing the production Docker image locally:

1. Copy the production environment file:

cp .env.production.example .env

1. Start the services:

docker compose -f docker-compose.production.yml up -d

The application will be available at http://localhost:8080.

To use a different port, set APP_PORT:

APP_PORT=3000 docker compose -f docker-compose.production.yml up -d

Whisper Money can be easily deployed to Coolify using our Docker Compose template.

1. In Coolify, create a new resource and select Docker Compose
2. Choose Empty Compose File as the source
3. Paste the contents from our template: 👉 whisper-money.yaml
4. Deploy!

The template includes:

• Whisper Money application container
• MySQL 8.0 database with health checks
• Persistent volumes for data and storage
• Auto-generated database credentials

Note: APP_KEY and APP_URL are auto-configured. The container generates an APP_KEY on first startup if not provided.

Connecting a bank — the flow that pulls a year of movements in and keeps them current by itself — goes through EnableBanking, a licensed open-banking aggregator. It is optional and a paid third-party service, so a self-hosted install ships with it switched off.

Without it the app runs perfectly well: accounts are created by hand or filled from a CSV import, and the brokers and exchanges (Indexa Capital, Coinbase, Binance, Bitpanda, Wise, Interactive Brokers) connect with their own API key and never touch EnableBanking. The bank connection is simply not offered anywhere in the UI.

To switch it on you need two things from EnableBanking: an application id, and an RSA key pair whose public half is registered against that application — every API call is a JWT signed with the private half (RS256, the application id as kid). The application's redirect URL has to be <APP_URL>/open-banking/callback, which is where the bank sends the user back.

All three are required together. With any of them missing, open banking stays off and nothing else is affected.

Whisper Money's AI features (transaction categorization and automation-rule suggestions) run on laravel/ai and default to Google Gemini. The provider is configurable independently of the model, so you can point the app at any text provider laravel/ai supports — gemini, openai, anthropic, azure, groq, xai, deepseek, mistral, a self-hosted Ollama server, or openai-compatible for any endpoint speaking the OpenAI API. Ollama is the headline case because it keeps AI processing fully local and private — data never leaves your infrastructure — but the switch is generic.

Each provider needs its own credentials configured for laravel/ai (e.g. GEMINI_API_KEY, OPENAI_API_KEY, ANTHROPIC_API_KEY, or OLLAMA_URL). An unknown or non-text provider fails fast when the AI feature runs.

AI_PROVIDER=ollama
OLLAMA_URL=http://ollama.example.local:11434
AI_SUGGESTIONS_MODEL=gemma3:12b
AI_CATEGORIZATION_MODEL=gemma3:12b

Make sure the model is pulled on the Ollama server first (ollama pull gemma3:12b). Any other provider follows the same pattern: set AI_PROVIDER, that provider's credentials, and the *_MODEL vars to one of its models. Gemini remains the default, so existing deployments are unaffected.

Plenty of services and local servers speak the OpenAI Chat Completions API without being OpenAI: router/gateway services such as OrcaRouter, local runtimes like LM Studio or vLLM, hosted inference like Together or Fireworks, a LiteLLM instance, or your own corporate proxy. Set AI_PROVIDER=openai-compatible and point the app at one of them.

• OPENAI_COMPATIBLE_URL is required — it is the base URL the app appends chat/completions to, so give it the versioned root (e.g. .../v1). Leaving it empty fails when the AI feature runs.
• OPENAI_COMPATIBLE_API_KEY is optional — when set it is sent as Authorization: Bearer <key>. Leave it empty for a local server that does not authenticate.
• The *_MODEL vars must name a model that endpoint serves. There is no default, and the Gemini defaults are meaningless to it.
• There is a single openai-compatible slot, so only one such endpoint can be configured at a time. To reach two of them, put a router in front.

Example with OrcaRouter:

AI_PROVIDER=openai-compatible
OPENAI_COMPATIBLE_URL=https://api.orcarouter.ai/v1
OPENAI_COMPATIBLE_API_KEY=sk-orca-...
AI_SUGGESTIONS_MODEL=orcarouter/auto
AI_CATEGORIZATION_MODEL=orcarouter/auto
AI_REPORTS_MODEL=orcarouter/auto

orcarouter/auto lets OrcaRouter pick the model; naming one directly (provider/model-name) works too. Any other OpenAI-compatible endpoint follows the same shape — only the URL and the model names change.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.