Lets show you...
How it works
Istina helps you make sense of the news in one place.
It pulls stories from different publishers, removes repeats, checks each article for potential bias and claim reliability, and then gives you a clear report that is easy to understand and use.
1. Ingest from trusted feeds
Istina fetches one or more RSS feeds, parses each entry into a clean article record, and keeps going even if one feed fails. Built-in retries and request timeouts help keep ingestion reliable in real-world network conditions.
2. Normalize and deduplicate
Every article gets a stable fingerprint based on source, URL, and publish time. That lets Istina detect duplicates consistently and avoid re-adding content you already have, while preserving a clean historical dataset.
3. Analyze only what is new
Istina selects unscored articles and runs them through a pluggable analysis provider
For each article, Istina produces…
Overall bias label: left, center, right, or unknown
Rhetorical manipulation flags
Claim-check objects with verdicts and confidence
A normalized confidence score and timestamp
If model output is malformed, Istina safely normalizes/falls back instead of breaking the pipeline.
4. Report with clarity
Istina can render:
Summary reports: total articles, analyzed count, source distribution, and bias-label distribution
Full reports: per-article detail with linked analysis output
1. Ingest from trusted feeds
Istina fetches one or more RSS feeds, parses each entry into a clean article record, and keeps going even if one feed fails. Built-in retries and request timeouts help keep ingestion reliable in real-world network conditions.
2. Normalize and deduplicate
Every article gets a stable fingerprint based on source, URL, and publish time. That lets Istina detect duplicates consistently and avoid re-adding content you already have, while preserving a clean historical dataset.
3. Analyze only what is new
Istina selects unscored articles and runs them through a pluggable analysis provider
For each article, Istina produces…
Overall bias label: left, center, right, or unknown
Rhetorical manipulation flags
Claim-check objects with verdicts and confidence
A normalized confidence score and timestamp
If model output is malformed, Istina safely normalizes/falls back instead of breaking the pipeline.
4. Report with clarity
Istina can render:
Summary reports: total articles, analyzed count, source distribution, and bias-label distribution
Full reports: per-article detail with linked analysis output
1. Ingest from trusted feeds
Istina fetches one or more RSS feeds, parses each entry into a clean article record, and keeps going even if one feed fails. Built-in retries and request timeouts help keep ingestion reliable in real-world network conditions.
2. Normalize and deduplicate
Every article gets a stable fingerprint based on source, URL, and publish time. That lets Istina detect duplicates consistently and avoid re-adding content you already have, while preserving a clean historical dataset.
3. Analyze only what is new
Istina selects unscored articles and runs them through a pluggable analysis provider
For each article, Istina produces…
Overall bias label: left, center, right, or unknown
Rhetorical manipulation flags
Claim-check objects with verdicts and confidence
A normalized confidence score and timestamp
If model output is malformed, Istina safely normalizes/falls back instead of breaking the pipeline.
4. Report with clarity
Istina can render:
Summary reports: total articles, analyzed count, source distribution, and bias-label distribution
Full reports: per-article detail with linked analysis output