Review Ingestion

SENTINEXT fetches public Steam reviews for a given AppID and stores them so you can analyze, search, and export them later.

Fetch From Steam

PHASE 01

We paginate Steam's reviews endpoint and collect review text plus useful metadata.

Reviews are pulled from the public Steam reviews API (no Steam login required). We store fields like language, recommendation, timestamps, helpful votes, and playtime so you can filter and drill down later.

Filters & Safety

PHASE 02

Rate limits, deduplication, and optional filters keep runs reliable.

The backend supports fetching recent, updated, best, or all reviews and can work across multiple languages. The hosted dashboard focuses on recent reviews and keeps limits in place to stay fast and stable.

Persistence Layer

PHASE 03

PostgreSQL storage for search, exports, and repeat analyses.

Data is persisted to PostgreSQL along with the AI labels and evidence quotes. If you analyze the same game again, cached labels can be reused instead of re-calling the model.

Defaults & Limits

Dashboard Default
1,000 REV/RUN
Backend Fetch Cap (Default)
5,000 REV
Fetching reviews...
74%
Review Distribution by Language
English
45%
Chinese
18%
Russian
12%
German
8%
Other
17%