Check an article
Paste the full article text below. Headline, source URL, and author are optional — leaving a field blank scores it as neutral, not suspicious.
Names and places are masked first. Then, the article’s writing style and source are reviewed.
Credibility report
"New curfew order tonight?" · analyzed just now
Where this came from
We check the website and any named author or organization against sources we already know are trustworthy or untrustworthy.
How it's written
We look for wording built to provoke strong emotions, and automatically hide personal names to protect people's privacy.
Settings
Your account, interface preferences, and how your data is handled.
Guest Session
You are browsing TunAI as a guest. All your check records, diagnostic scores, and reports are saved locally on this computer and browser. No account registration is required.
Clear local data
Clears every check stored in this browser. This can't be undone.
Interface language
Changes menus, buttons, and labels only. Article text can always be entered in English, Tagalog, or Taglish — that never has to match.
Privacy safeguards
How TunAI protects people named in the articles you check.
Your data
Flagged Articles
Review flagged articles and submit verification notes.
No flagged articles pending review
All suspicious submissions have been audited.
| Domain | Standing | Incidents | Audit Notes | Last Updated | Actions |
|---|
No sources found
Try clearing your filter or add a new news source.
| Name / Entity | Type | Standing | Risk Cushion | Audit Notes | Last Updated | Actions |
|---|
No authors or organizations found
Try clearing your filters or add a new entry.
Help Center & User Guide
Learn how TunAI analyzes news content, verifies credibility signals, and protects your privacy.
How to Use TunAI
A practical step-by-step walkthrough of the credibility check workflow.
Open TunAI
Open the website and accept the academic research disclaimer. You can use the public checking features without creating an account.
Enter the article
Paste the news article's text into the analysis field, or provide its URL if you want TunAI to retrieve the article. Use the input method available on the page.
Check the input
If using a URL, make sure it is correct and points to an accessible article. If automatic retrieval fails, try pasting the article text directly.
Run the analysis
Start the credibility check and wait for the system to process the submitted content. Processing time may vary depending on the input and retrieval services.
Review the results
Read the displayed risk classification, model score, and available domain or author credibility indicators. Treat these as signals for further evaluation, not a final verdict.
Explore the explanations
Review available text highlights and explanation tools, including LIME or SHAP when provided. These help explain which features influenced the model's prediction; they do not independently prove whether a claim is true.
Save or review your results
Use the available report download and history features to revisit your checks. Guest history is stored in your browser, so clearing browser data or switching devices may affect whether those records remain available.
Frequently Asked Questions
Common questions about detection, scoring, privacy, and technology.
What is TunAI?
TunAI is an academic research system designed to analyze news articles and provide indicators related to their credibility. It combines text-based analysis with available domain and author credibility signals to help users assess news content.
Does TunAI determine whether an article is definitely fake or real?
No. Its output is a model-based assessment, not definitive proof. A low-risk result does not guarantee that an article is true, and a high-risk result does not independently prove that it is fake. Always verify important claims using reliable sources.
How can I analyze an article?
You can submit article text through the text input or provide an article URL if URL analysis is available. If the system cannot retrieve a webpage, try copying and pasting the article text instead.
What do the risk classifications mean?
Low Risk indicates a lower level of detected risk according to the system's assessment. Uncertain indicates that the result may be inconclusive or needs further examination. High Risk indicates a higher level of detected risk. These classifications are system outputs, not guarantees of truth or falsity.
What does the model score mean?
The displayed score represents the model's assessment according to the scoring method implemented in TunAI. It should not automatically be interpreted as the exact probability that an article is true or false unless the system explicitly defines and validates it that way.
Why does TunAI examine the domain and author?
Domain and author indicators provide additional context alongside the article text. They can help users consider the source of a claim, but a reputable source can publish an inaccurate article, and an unfamiliar source can publish accurate information.
What are LIME and SHAP explanations?
LIME and SHAP are explainability methods that can help show which input features influenced a model's prediction. Their visualizations describe the model's behavior; they are not independent fact-checking or proof that a claim is false.
Which languages does TunAI support?
TunAI is designed to support English and Filipino/Tagalog content, including Taglish where supported by the model. Performance may vary by language, writing style, topic, and the quality of the submitted text.
Does TunAI save my analysis history?
For guest use, the website stores analysis history in your browser's local storage. This allows you to revisit records in the same browser. Records may not appear on another device or browser, and clearing site data may remove them.
What happens to the article or URL I submit?
The submitted content is processed by TunAI to perform the requested analysis. URL retrieval and other enabled features may involve external services. Avoid submitting confidential or sensitive information, and consult the Privacy Policy for the system's actual data-handling details.
What should I do if the result seems incorrect or the analysis fails?
Check that the article text is complete and the URL is accessible. If retrieval fails, try pasting the text directly. Compare the article with multiple reliable sources and original documents. If TunAI provides a feedback or reporting feature, use it to report a problem; otherwise, follow the contact instructions provided by the project team.
