Attaching documents

An explainer gets much better when it can teach from your actual material instead of a general example. On scrimba.com/explain, attach files to your prompt and the lesson is built from them.

What you can attach

• Code files: .js, .ts, .py, .html, .css, .json, and other source files
• Text and markdown: notes, READMEs, logs, config files
• PDFs: papers, documentation, slides
• Images: screenshots, diagrams, photos of a whiteboard
• Jupyter notebooks: .ipynb files

Drop files onto the composer, or pick them from the attach button. A long paste does the same thing: paste a big chunk of code or text into the prompt and it becomes an attachment instead of flooding the input.

What to use it for

• A source file. ›walk me through this file, function by function
• A log. ›explain what went wrong in this log
• A PDF paper. ›teach me this paper's main idea
• A screenshot of an error. ›what is this error and how do I fix it
• A notebook. ›explain what this analysis does, cell by cell

The explainer quotes your real code and data on its slides, with the narration pointing into it, rather than teaching around it. Attached source turns into code slides; a paper or a set of notes leans on diagrams and images instead.

If the material you want taught is a repository rather than a file, a connected coding agent reads it directly and skips the attaching.