Overview
UCMDB Support AI Chatbot is an AI-powered support assistant for OpenText UCMDB, described as a personal project. It uses Retrieval-Augmented Generation (RAG) to answer support questions by searching a local knowledge base of KCS articles, support incidents, and UCMDB documentation, then generating answers with a local LLM via Ollama. The project states that no data leaves the machine it runs on and all models run locally.
The retrieval pipeline converts a user question to a vector with all-MiniLM-L6-v2, searches a ChromaDB vector store using cosine similarity for top 20 candidate chunks, reranks with cross-encoder/ms-marco-MiniLM-L-6-v2 to top 5 chunks, and uses Ollama LLM (Mistral 7B) to generate an answer grounded in retrieved chunks. A FastAPI + HTML/JS frontend displays the answer and collapsible source citations.
Knowledge sources include KCS Knowledge Articles (.txt scraped from portal.microfocus.com, about 730 articles), R&D Incident Reports (.txt, over 28,000 incidents), and UCMDB Documentation (.pdf, multiple volumes), totaling 52,820 chunks in the vector store. The repository excludes knowledge-base directories because they contain proprietary support data, so users must populate those folders with their own data.
Quick start prerequisites include Python 3.10+, Ollama installed and running as a service, the Mistral model pulled, and CUDA-enabled PyTorch for GPU acceleration on embedding. The chatbot exposes API endpoints for the web UI, chat, health check, and debug retrieval, and configuration settings control the Ollama model, candidate and final chunk counts, max tokens, and minimum rerank score.
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Support teams use UCMDB Support AI Chatbot to answer OpenText UCMDB support questions by searching a local knowledge base of KCS articles, incidents, and documentation.
- Developers use it to run a local FastAPI chatbot server with a web UI.
- Administrators use it to refresh the knowledge base by dropping new .txt or .pdf files into the appropriate folder and rebuilding the vector store.
- Operators use it to debug retrieval by viewing all 20 candidates with cosine distances and rerank scores.
- Users use it to view answers with collapsible source citations in a web UI.
- Teams use it to keep support data local because no data leaves the machine it runs on.
Features
AI assistant
Natural language interface
Workflow automation
Reviews
No reviews yet. Be the first to share your experience!