Checking News Before Sharing: The WizeUp Workflow

Written by the Fulldive product engineering team, based on direct inspection of the repositories listed below.

This post is a practical walk-through of one job: how to check if news is fake before sharing, on an Android phone, using WizeUp. It is a use-case post, not a product tour. If you want to check news before sharing — especially claims you saw on social media or in a group chat — this is the workflow we designed the app around, traced back to the commits that shipped each piece.

For the honest limits of AI-assisted claim review, read what an AI fake news checker can and cannot do alongside this post. The workflow only works if you understand what AI Chat can and cannot do.

The situation

Someone forwards you a post. A screenshot of a headline. A short video claiming something happened. You are about to share it to a group, or reply. You feel a small hesitation. That hesitation is the moment WizeUp is designed for.

The workflow below takes about two to three minutes and is meant to raise your confidence enough to share, hold, or ask a question, without pretending to settle the truth of the claim on its own.

Step 1 — Open the trend in WizeUp

If the story is circulating at all, it is likely on WizeUp’s trends surface. Open the app. The home feed, rebuilt in commit 11fc50e4cf for version 6.2.0, shows what is trending. If it is a fast-breaking story, rising trends (covered in personalized feeds, trends, and daily digest) will surface it even before the top trends list does.

Tap the trend. The trend detail screen opens. This is the surface where everything else happens.

If you know the direct URL, the wizeup:// deep-link router added in commit f9cce9d25a can take you straight to the trend — more on that in deep links and App Info in WizeUp.

Step 2 — Identify the specific claim

Most viral posts mix a real fact, an opinion, and a framing. The first thing to do is separate them. AI Chat — refactored in commit 3e90b710c3 — is scoped to the trend in front of you. Ask it, plainly: “What is the specific factual claim this post is making?”

A good reply will do three things:

Name the claim as a checkable proposition (“X organisation said Y on Z date”).
Separate it from the framing (“the post implies this is unusual”).
Flag any parts that are opinion rather than fact.

If AI Chat returns something vague, ask a sharper follow-up. The purpose at this stage is not to verify anything — it is to know what needs verifying.

Step 3 — Ask where to check

Next question to AI Chat: “What primary sources would I check this against?”

Useful answers point to types of sources, not just names:

Official statements or press releases, if the claim is about an organisation.
Filings or datasets, if the claim is numerical.
The original video or full transcript, if the claim is about a quote.
Reputable outlets that have covered the topic, as secondary verification.

Treat the reply as a list of doors to open, not as a verdict. The IFCN code of principles — the baseline that professional fact-checkers work to — is built around transparency of sources, and that is the standard your own review should try to imitate.

Step 4 — Open at least one primary source directly

This is the step that matters most. Open a primary source in WizeUp’s browser view directly. Read the actual quote, statement, or dataset. Do not rely on AI Chat’s summary of it.

If Google’s Fact Check Tools already index a published fact-check on the claim, that is a strong signal — but only if the published check is recent and addresses the same claim, not a lookalike.

Step 5 — Check how the claim is being framed

Back on the trend detail page (d328d2d8a5), look at comments (bcf547c606). Not to defer to them, but to see what angles other readers are raising. If several comments independently note the same counter-context — a missing date, a clipped video, a source retraction — pay attention.

Comments are not evidence. They are signal about where to look next.

Step 6 — Decide, with a caveat if needed

At this point you have:

The specific claim, clearly stated.
At least one primary source you read yourself.
A rough sense of how the claim is framed in the wild.

Decide whether to share, hold, or ask a question. If you share, add what you actually know: “the filing says X, which is narrower than the headline.” That phrasing is more useful to everyone in the chat than either silent forwarding or a blanket “fake news” dismissal.

How the app supports the workflow, in repo terms

The commits behind this workflow, for readers who want to trace it:

App Info screen (commit c0bb2c6c3f) — a dedicated surface that makes WizeUp’s purpose, limits, and support links visible from inside the app. Covered in deep links and App Info in WizeUp.
Home feed and trend detail analytics (commit 0536b97ea8) — instrumentation that lets us see whether this workflow is actually being used. We treat those events as product instrumentation, not a behavioural profile we sell.
AI Chat refactor (commit 3e90b710c3) — the surface that makes steps 2 and 3 possible.

Limits

Since this is a use-case post about a sensitive topic, the limits are not optional. They are the post.

WizeUp is not a guaranteed claim checking app. No mobile app is. Following this workflow makes you more likely to catch a bad claim. It does not guarantee it.
AI Chat can be confidently wrong. Every step that involves AI Chat is intended to generate leads, not verdicts. The full rationale is in what an AI fake news checker can and cannot do.
A claim that is trivially true in one locale can be misleading in another, and vice versa. If the claim hinges on local context, try to find a local primary source.
WizeUp is not a fully private browser. We do not claim it is. System-level ad and tracker blocking is a separate tool; see Wize AdBlock VPN.
If a claim is important enough to share, it is important enough to be willing to not share.

Where this fits

The broader product history — why an Android browser codebase turned into a claim-review tool at all — is in how Fulldive Browser became WizeUp. The engineering work that makes this workflow feel fast is in reducing ANR and cold start in a browser. The broader Fulldive ecosystem is in the company story and the projects list.

You can install WizeUp from /project/fulldive-browser/. Questions about the workflow are welcome at support@fulldive.com, and the FAQ covers cross-product questions.

Sources

Last updated: 2026-04-16. Commit hashes and version numbers are drawn from Fulldive repositories inspected on 2026-04-13.