• Installed ocm-cli
• Logged in (ocm login --use-auth-code)
• Python 3.12
• Installed uv for Python package management (pip install uv)

To use the summarization code, you will need an OpenAI-compatible API. The current code is written assuming LM Studio is running and Gemma 3 27b is loaded. If you are using ollama or another service, change the OpenAIModel configuration in main.py.

Depending on how much text was added to each of the events, the context length may need to be raised to accomodate. I run with 8096.

These variables must suit your environment. Examples are provided here for what I have most recently used.

export OPENAI_API_KEY="lm-studio" OPENAI_BASE_URL="http://127.0.0.1:1234/v1" OCM_TOKEN=$(ocm token)
export SUMMARIZATION_MODEL_NAME=mistral-small-3.1-24b-instruct-2503 EDITOR_MODEL_NAME=mistral-small-3.1-24b-instruct-2503

The reports and summaries can be generated easily in your shell.

uv run reports.py
uv run summaries.py

If you wish to print to PDF or view the reports in your browser, launch the web interface with Streamlit.

uv run streamlit run Incident\ Report.py

The reports and summaries can be found in ./incident-reports and ./incident-summaries