Inspiration Cybersecurity investigations are often slow, manual, and dependent on highly experienced analysts. We wanted to build a system that could automate evidence analysis, correlate findings across multiple sources, and help security teams respond faster and more accurately.
What it does Protocol SIFT is an autonomous cyber investigation platform that ingests forensic evidence, extracts entities, reconstructs attack relationships, detects threats, builds timelines, and generates explainable incident reports from a single workspace.
How we built it We built Protocol SIFT using a React-based frontend and a FastAPI backend. MongoDB stores investigations, evidence, entities, findings, timelines, and reports. The platform processes uploaded logs, normalizes data, extracts entities, builds relationship graphs, generates findings, and visualizes the investigation through interactive dashboards.
Challenges we ran into Correlating data from different evidence sources
Designing a scalable investigation workflow
Building graph relationships without dedicated graph databases
Creating explainable findings instead of black-box results
Ensuring the frontend could visualize complex investigation data clearly
Accomplishments that we're proud of Built an end-to-end investigation pipeline
Automated entity extraction and threat detection
Generated attack timelines and investigation reports
Developed an interactive knowledge graph visualization
Created a professional SOC-style investigation dashboard
What we learned The importance of data normalization in cybersecurity
How graph-based relationships improve investigations
Designing explainable AI workflows for security operations
Building scalable full-stack systems for real-world incident response
What's next for Protocol SIFT Integrate forensic tools such as Volatility, Sigma, YARA, Zeek, and Chainsaw
Add AI-powered Investigator, Verifier, and Skeptic agents
Implement advanced confidence scoring and self-validation
Integrate MITRE ATT&CK mapping and threat intelligence feeds
Support real-time investigations and collaborative SOC workflows
Evolve into a fully autonomous cyber incident response platform capable of reducing investigation time from hours to minutes.
Built With
- fastapi
- git
- javascript
- langgraph
- mongodb
- pydantic
- python
- react
- reactflow
- restapi
- tailwindcss
- typescript
- vite
- websockets
Log in or sign up for Devpost to join the conversation.