Step 1
Watch the clip more than once, first without the caption.
Session mode surfaces the fastest pathways for live workshop use.
Practical workflow
Separate what the footage actually shows from what the caption or claimed context adds to it.
A short clip is paired with a strong caption that claims more than the footage clearly shows.
Watch the clip more than once, first without the caption.
Write down what is directly observable before reading the interpretation.
Use AI to generate questions about place, time, and context.
Extract frames or use tools such as InVID.
Search for a longer or original version if it exists.
Compare the caption with what the clip actually shows.
Decide whether the claim is supported, overstated, or still unconfirmed.
This guide organizes tool choice around the task: breaking a claim down, reviewing documents, tracing an archived page, or supporting image and video checks.
Four short workflows for common cases: a textual claim, a suspicious image, a clip with unclear context, and a document set that needs structure and checking.