Classify content
Submit Markdown text and receive matching content-classification labels. Long requests can be queued and fetched later with the job UID. The classification can also be validated and justified by the LLM when needed.
AIPITCH API service
Topic classification.
API service for classifying content with the MISP content-classification taxonomy through an Ollama-backed AI engine. This project is part of AIPITCH, an international consortium focused on AI-based tools for cybersecurity operations, founded by the European Commission.
Checking health
- Taxonomy
- MISP content-classification
- Open Source Project
- Topic Classifier on GitHub
- Health
- Waiting for scheduler probe
Qualification
Classification accuracy is tested across different sources such as forums and chat content, with techniques to avoid label multiplication and keep the results well qualified.
Classification
The classification engine runs on premise and scores content against the MISP content-classification taxonomy. It can be configured to use different open source models depending on the deployment and available compute. Taxonomy JSON
Who is CIRCL?
CIRCL is the Computer Incident Response Center Luxembourg. It is a public CSIRT focused on incident response, threat intelligence, and security support for organisations. It also contributes to open source to share tools and methods with the security community.
Access on request
This Topic Classifier service is operated by CIRCL and offered on request. Access can be granted by contacting info@circl.lu.
What it fits
It can be used in surveillance pipelines, CTI workflows, and other security automation chains to classify security-related Markdown content against the MISP content-classification taxonomy using an Ollama-backed AI engine.