Analogy generation prompt builder
The fastest way to explain something hard is to connect it to something the listener already understands. A good analogy does the heavy lifting — but a lazy one (“it’s like a highway, sort of”) teaches nothing and a wrong one teaches the wrong thing. This builder produces a prompt that asks an LLM to generate analogies tuned to a specific audience and source domain, with an explicit structural mapping so the comparison genuinely transfers understanding.
How it works
You enter the concept, choose the target audience, optionally pick a source domain to draw the analogy from, and set the mapping depth. The tool composes a prompt that tells the model to first identify the essential mechanism of the concept, then find a source domain the audience knows well, then map each part of the concept onto a part of the analogy. At deeper settings the prompt also requires the model to state where the analogy breaks down — the single most useful sentence in any explanation, because it stops the reader from trusting the comparison too far.
Tips and example
Match the source domain to your audience’s life, not your own — a plumbing analogy lands with a contractor and bounces off a teenager. Use the mapped or deep format whenever the concept has moving parts; the part-by-part alignment is where the actual learning happens. A worked example: concept “database indexing,” audience “junior developers,” domain “library,” depth “deep” — the model maps the index to a card catalog, the lookup to finding a shelf, and then notes the analogy breaks down because rebuilding an index has a write cost a catalog does not. Generate two or three and keep the one that clicks.