Intelligence
Pipeline.
From Steam reviews to structured insights, with evidence.
1. Search & Ingest
Search a game (name or App ID) and run an analysis. SENTINEXT fetches recent Steam reviews and saves results to your local database.
- > Search by name or paste an App ID.
- > Fetch recent reviews from Steam (public endpoint; no Steam login).
- > Runs in the background with progress tracking and PostgreSQL persistence.
2. Classification
AI labels each review with a consistent taxonomy and extracts evidence quotes.
- > Assigns 1-6 tags (e.g., technical/performance, ui_ux_accessibility/quality_of_life).
- > Separates issues vs feature requests.
- > Stores verbatim evidence quotes for every tag so you can audit results.
"Game keeps crashing mid-battle since the last patch, and the inventory menus are impossible to navigate with a controller."
"keeps crashing mid-battle"
-> stability_crashes
"inventory menus impossible to navigate with controller"
-> menus_hud
3. Quantified Insights
Dashboards, review explorer, AI agent chat, and comparisons: go from numbers to decisions with evidence and drill-down.
- > Dashboard: category breakdown, top issues/requests, and trends.
- > AI agent chat: questions and charts grounded in your analyzed data.
- > Compare games, export CSV/JSONL, and generate monthly PDF reports.
Recommendation
87%
Issue Rate
23%
Request Rate
15%
Reviews
1,000
4. Agent Chat
Ask natural-language questions about your analyzed data. The agent answers with charts, evidence, and quotes from real reviews.
- > Natural-language questions -- no queries to write.
- > Answers with inline charts and interactive breakdowns.
- > Verbatim review quotes as auditable evidence.
What are the top technical issues?
Across 1,000 reviews, the most reported technical issues are:
Evidence -- stability
"Game crashes every time I enter the third act, lost hours of progress."