QATIS

Turnkey OSINT collection and AI-assisted evaluation. Collect from web, academic, and optional social sources; export, dedupe, and score with an LLM; review structured outputs.

Screenshot

Features

• Multi-source collection: Google (via ScraperAPI), OpenAlex (academic), optional social (YouTube, Reddit, VK)
• AI evaluation: intel/non-intel, confidence, rationale, PMESII tags, Admiralty codes
• Outputs: CSVs, PMESII markdown summary, BibTeX
• Interfaces: Streamlit GUI and CLI

Requirements

• Python 3.11+ for GUI (CLI works on newer Python but GUI is tested on 3.11)
• Keys: OPENAI_API_KEY, SCRAPERAPI_API_KEY (trial available)
• Optional social credentials: Reddit, VK

Installation

Recommended (macOS):

brew install python@3.11
cd /Users/emilwl/Documents/QATIS
/opt/homebrew/bin/python3.11 -m venv .venv
source .venv/bin/activate
pip install -e .
pip install -e '.[ui]'

CLI-only (newer Python):

pip install -e . --break-system-packages

Configure keys

qatis configure
# Provide OPENAI_API_KEY and SCRAPERAPI_API_KEY

Run (GUI)

./start_ui.sh
# or
qatis ui
# Opens at http://localhost:8501

Run (CLI)

# One-shot pipeline
qatis run-all --queries queries.yaml --top-k 10

# Or step-by-step
qatis collect --queries queries.yaml --top-k 20
qatis export --results-dir search_results/<timestamp>
qatis analyze --results-dir search_results/<timestamp>

Example queries

areas:
  - 'Moldova infrastructure 2024'
structures:
  - 'Moldelectrica ownership filetype:pdf'

Optional social:

youtube:
  queries:
    - 'Moldova energy 2024'
reddit:
  subreddits:
    - 'europe'
  queries:
    - 'Moldova infrastructure'

Outputs

• results_scored.csv — All sources with labels, confidence, PMESII, Admiralty, rationale
• results_scored_intel.csv / results_scored_non_intel.csv
• pmesii_infrastructure.md — PMESII-organized summary
• intel_sources.bib — BibTeX of intel sources
• analysis_cache.jsonl — Cache to avoid re-analyzing

Project structure

qatis/
  collect/        # collectors for web/social
  analyze/        # analysis pipeline
  export/         # exporters
  ui_app.py       # Streamlit GUI
start_ui.sh       # GUI launcher
queries.yaml      # example queries
search_results/   # outputs by timestamp

Troubleshooting

• Missing keys: run qatis configure
• UI won’t start on Python 3.14: use Python 3.11
• Empty results: check ScraperAPI balance, simplify queries
• Social scraping failures: ensure optional deps, note that X/Twitter is often blocked

License

MIT