- A queue system using BullMQ (which uses Redis)
- Analyze every queue message with an AI LLM (e.g. with Gemini or OpenAI)
- Log all requests including the AI output
- Add extra custom logic for the processed message
- Database integration (MongoDB by default)
- Dockerfile and Compose files for easy deployments
- Modular and reusable framework so you can deploy multiple AI queue microservices
- Copy the
.env.examplefile to.envand fill in the necessary variables - Populate either the Gemini or OpenAI API keys
- Find all and replace the word "finallyqueue" with your project name
- Update ./src/prompts/input-prompt.txt
- Update ./src/prompts/output-schema.ts
npm run dockerto run the network in dockernpm run testto send a test message- view logs in the queue container or the logs collection in the db (port 27017 by default)