An intelligent assistant that automates GitHub issue triage using AI, Redis Stack, and Spring Boot.
This project was built for AI Week @ Redis to streamline the process of managing and labeling GitHub issues across client libraries like Jedis and Lettuce.
- 🏷️ Auto-labels new GitHub issues using LLMs (GPT-4, Claude, Gemini via LiteLLM)
- 🔍 Finds similar past issues using Redis vector search (RediSearch + HNSW)
- 🔔 Sends Slack notifications to relevant maintainers
- 💬 (Optional) Posts enriched triage results as GitHub comments
- 🧠 (Planned) Stack trace root cause analysis
- 🧪 (Planned) Test coverage gap detection
| Component | Tool/Service |
|---|---|
| Backend | Java 24, Spring Boot 3.5 |
| LLM Integration | LiteLLM proxy to GPT-4 / Claude / Gemini |
| Redis Vector DB | Redis |
| Redis Client | redis/jedis |
| GitHub | Webhook API + GitHub Java API |
| Slack | Incoming Webhooks |
| Testing | JUnit 5, Testcontainers, Mockito |
[ GitHub Webhook ]
|
v
[ Spring Boot API ]
├── LabelingService → calls LiteLLM
├── RedisVectorStoreService → stores issue vectors in Redis
├── SemanticSearchService → finds similar issues
└── SlackNotifier → sends Slack alerts
- Java 21 or 24
- Maven 3.9+
- Docker (for Redis Stack via Testcontainers)
- Slack Webhook URL
- LiteLLM endpoint + API key (for OpenAI, Claude, or Gemini)
git clone https://github.com/your-org/ai-triage-assistant.git
cd ai-triage-assistant
./mvnw clean installIntegration tests spin up a full Redis Stack container:
./mvnw test- Auto-labeling via LLM
- Redis vector similarity search
- Slack integration
- Testcontainers-based integration tests
- Root cause analysis from stack traces
- Test coverage detection
- GitHub comment bot