Read before this
Read these direct prerequisites in order. Each guide lists any earlier context it assumes.
Ways to add a source
- Paste a public URL into the prompt.
- Use the plus button to upload supported documents, slides, images, audio, or video.
- Use Screen Recorder when the source is a product walkthrough you need to capture now.
- Reuse approved assets from your Studio uploads or Brand Kit when the picker offers them.
Tell ngram how to use it
An attachment alone does not explain your intent. Say whether it is the factual source of truth, a visual reference, a narration source, a brand asset, or a recording to polish.
For several sources, name the priority. For example: use the release notes for facts, the deck for structure, and the screenshots only for product visuals.
Prepare files for reliable results
- Use selectable text in PDFs and decks when possible.
- Put screenshots in the order the workflow happens.
- Trim unrelated pages or recordings before upload.
- Use clear filenames and mention them in the prompt.
- Do not upload secrets, personal data, or material you do not have permission to process.
Review or remove attachments
- Open the current conversation and expand its attachment tray.
- Choose Edit when you need to manage uploaded items.
- Mark the files that are no longer needed, then save the change.
- If a new instruction still depends on a removed file, attach it again or paste the approved facts needed for that request.
When a source cannot be read
A private or blocked webpage may not expose its content to ngram. Export it to a supported file, paste the approved text, or provide screenshots instead. If a large source produces a vague script, narrow the request to the exact section or outcome you need.
Keep going
Write prompts that produce stronger scripts, visuals, pacing, and revisions.
Use enhanced visual analysis for websites, PDFs, images, and other visually important sources.
Use the Screen Recorder, browser permissions, recording tips, review, and editor handoff.