Gemini Prompts

Turn any assistant into a decisive, no-padding expert. Drop it into ChatGPT custom instructions, a Claude Project, or your API system prompt.

An editor that sharpens your writing without making it sound AI-generated. Works as ChatGPT custom instructions or a Claude system prompt.

Turns any assistant into a tutor that actually teaches — questions, hints, and understanding checks instead of answer dumps.

For developers: forces schema-exact, parseable output every call — for apps, agents, and data pipelines built on the API.

Turns Gemini into a research analyst — structured briefing, key players, what's debated, cited sources. Plays to Gemini's long context and Google grounding.

Repair, upscale, and colorize damaged or B&W photos using a 3-pass technique (restore → colorize → enhance) with decade-specific color palettes so the result looks historically authentic, not generically 'colorized.'

Play to Gemini's strongest capability — paste any visual (chart, screenshot, document, photo, or two images to compare) and get a type-specific structured analysis that extracts what actually matters, not just a description.

Built for Gemini's massive context window — paste multiple files, the full stack trace, and get root-cause analysis that sees architectural issues, not just line-level bugs. Includes the 'write a test that catches this' output.

Turns Gemini into an adaptive study coach that explains, quizzes, schedules review using spaced repetition, verifies understanding with the 'teach it back' method, and runs timed exam simulations with scoring.

Paste a messy table and get a full analysis pipeline — data cleaning first, pivot tables, statistical significance checks, the exact Sheets formulas, and a specific chart recommendation with the reasoning for why that chart type (not another) tells the story.

Paste a mediocre prompt and get a full diagnosis of its failure modes, a rewritten version with every improvement annotated, and a side-by-side scored comparison so you can see exactly what changed and why it's better.

Get a complete, deployable system prompt for a Custom GPT, Gemini Gem, or Claude Project — including the sections most people forget: personality guardrails, the 'don't' list, edge case handling, refusal patterns, and token budget management.

For complex work, the AI designs a sequence of chained prompts — each step feeds the next — instead of one overloaded prompt.

Turns an exported keyword list into clustered topic maps with intent labels, an opportunity ranking, and a content-gap shortlist — decision-ready, not a raw spreadsheet.

Produces a clear, balanced freelance contract clause checklist (scope, payment, IP, kill fee, liability) with fill-in fields — a discussion draft for attorney review, not a signable contract.

Turns AI into a disciplined sports-prop research assistant that surfaces relevant factors, flags uncertainty, and enforces bankroll discipline — educational, with explicit no-guarantee guardrails. It never tells you what to stake.

Creates an individualized, periodized powerlifting block (volume -> intensity -> peak/deload) tuned to your lifts, recovery, and equipment — educational, not medical advice.

Generates a structured, role-specific interview question bank with signal-rubric scoring and bias guardrails — for interviewers, not candidates cramming.

Converts a fuzzy feature idea into a structured PRD — problem, users, success metrics, scope, risks, open questions — ready for stakeholder review.

An interactive prompt that makes AI lead a reflective journaling session with a student — one age-appropriate question at a time, with privacy and supervision guardrails. (An AI prompt, not a printable question list.)

An interactive prompt that makes AI act as a warm first-date conversation coach — surfacing good questions, reading the vibe, and easing nerves. (An AI prompt, not a question list to recite.)

Produces a hardware-aware Ollama model recommendation for your task plus a ready-to-run install and run script with VRAM checks, instead of guessing a model name and hoping it fits.

Produces a reusable schema-validation and dtype-coercion script that flags every column that drifted from spec, coerces what it safely can, and quarantines what it cannot instead of producing silent NaNs.

Takes a report-style spreadsheet (merged headers, multi-row titles, wide months-as-columns) and reshapes it into a tidy long DataFrame with a documented step-by-step transform you can re-run on the next export.