OpenDeRisk is an AI-Native Risk Intelligence System providing comprehensive 24/7 application system risk management. It utilizes multi-agent collaboration to deliver in-depth root cause analysis and intelligent protection.
One-Minute Overview#
OpenDeRisk is an AI-driven risk intelligence management system designed for Site Reliability Engineering (SRE) teams, providing comprehensive application system protection. Through its multi-agent architecture, it quickly identifies root causes, visualizes diagnostic processes, and delivers in-depth analysis. Ideal for technical teams requiring automated problem diagnosis and risk assessment.
Core Value: Uses AI agent collaboration for root cause analysis, reducing traditional troubleshooting time from days to minutes.
Quick Start#
Installation Difficulty: High - Requires handling 20GB+ datasets, configuring multiple components, and understanding AI/SRE concepts
# Install uv package manager
curl -LsSf https://astral.sh/uv/install.sh | sh
# Install dependencies
uv sync --all-packages --frozen \
--extra "base" \
--extra "proxy_openai" \
--extra "rag" \
--extra "storage_chromadb" \
--extra "derisks" \
--extra "storage_oss2" \
--extra "client" \
--extra "ext_base"
# Download OpenRCA dataset
gdown https://drive.google.com/uc?id=1cyOKpqyAP4fy-QiJ6a_cKuwR7D46zyVe
# Start the service
uv run python packages/derisk-app/src/derisk_app/derisk_server.py --config configs/derisk-proxy-aliyun.toml
Is this suitable for me?
- ✅ Enterprise SRE Teams: Large technical teams needing automated root cause analysis and risk assessment
- ✅ Complex System Monitoring: Diagnosing issues in distributed systems and microservice architectures
- ❌ Small Projects: May be overly complex for resource-constrained small projects
- ❌ AI Beginners: Teams unfamiliar with AI/LLM concepts may struggle with configuration and maintenance
Core Capabilities#
1. DeepResearch RCA - Root Cause Analysis#
- Quickly locate problem origins through in-depth analysis of logs, traces, and code Actual Value: Reduces troubleshooting from days to minutes, significantly decreasing system downtime
2. Visualized Evidence Chain#
- Fully visualize diagnostic processes and evidence chains for transparent analysis Actual Value: Enables teams to quickly understand diagnostic result accuracy, improving confidence and efficiency in issue resolution
3. Multi-Agent Collaboration#
- Collaboration among SRE-Agent, Code-Agent, ReportAgent, Vis-Agent, and Data-Agent Actual Value: Specialized AI agents handle different aspects, providing more comprehensive and accurate analysis results
4. Open Source Architecture#
- Completely open and open-source architecture, ready to use in open-source projects Actual Value: Highly customizable with team-specific extensions, avoiding vendor lock-in
Technology Stack & Integration#
Development Language: Python Main Dependencies: uv package manager, OpenAI proxy, RAG, ChromaDB storage, OSS2 storage, MCP services Integration Method: API / SDK / Library
Maintenance Status#
- Development Activity: Actively developed with v0.2 released and a clear roadmap
- Recent Updates: Version 0.2 released in October 2025
- Community Response: Has community groups (DingTalk) and acknowledges related projects like DB-GPT and MetaGPT
Documentation & Learning Resources#
- Documentation Quality: Comprehensive
- Official Documentation: OpenDerisk Documents and DeepWiki available in the repository
- Example Code: Quick start tutorials and scenario demonstrations provided