This is a project that allows you to categorize uncategorized transactions for Actual Budget using OpenAI or OpenAI specification compatible API, like a self-hosted Ollama server.
The app sends requests to the OpenAI API to classify transactions based on their description, amount, and notes.
❌ When a transaction cannot be classified, it is marked in Notes as "not guessed," and it will not be classified again.
Sample docker-compose.yml
file:
services:
actual_server:
image: docker.io/actualbudget/actual-server:latest
ports:
- '5006:5006'
volumes:
- ./actual-data:/data
restart: unless-stopped
actual-ai:
image: docker.io/sakowicz/actual-ai:latest
restart: unless-stopped
environment:
- ACTUAL_SERVER_URL=http://actual_server:5006
- ACTUAL_PASSWORD=your_actual_password
- ACTUAL_BUDGET_ID=your_actual_budget_id # This is the ID from Settings → Show advanced settings → Sync ID
- CLASSIFICATION_SCHEDULE_CRON=0 */4 * * * # How often to run classification.
- CLASSIFY_ON_STARTUP=true # Whether to classify transactions on startup (don't wait for cron schedule)
- SYNC_ACCOUNTS_BEFORE_CLASSIFY=false # Whether to sync accounts before classification
- OPENAI_API_KEY=your_openai_api_key
# - OPENAI_MODEL= # optional. required if you want to use a specific model, default is "gpt-3.5-turbo-instruct"
# - OPENAI_BASE_URL= # optional. required if you don't want to use the OpenAI API but OpenAI compatible API, ex: "http://ollama:11424/v1
# - ACTUAL_E2E_PASSWORD= # optional. required if you have E2E encryption
# - NODE_TLS_REJECT_UNAUTHORIZED=0 # optional. required if you have trouble connecting to Actual server
I'm not a Node developer. I have no experience with AI or GitHub actions. I've created this script to help categorise hundreds of transactions in my Actual Budget and decided to publish it.
Feel free to suggest changes or open a MR.