The Gemini Prompts Library
Google Gemini's 1M+ token context and native multimodality demand a different prompt style. This library is built specifically for it.
Gemini rewards explicit structure, role assignment, and long-context grounding. Every prompt here is tuned for Gemini 2.0 Pro, Gemini 2.0 Flash, and Gemini 1.5 Pro, with notes on which variant gives the best result.
Use these for long-document analysis, multimodal reasoning over screenshots and PDFs, codebase review, RAG pipelines, and Google Workspace automation.
Featured loops
The Accountability Bot
Drafts a sequence of check-in messages that increase in urgency without being annoying.
The Authority Signal Loop
Injects specific proof and lived experience to differentiate from generic AI content.
The API Polish Loop
Iteratively refines API endpoints for better developer experience, naming consistency, and response efficiency.
The Objection-Crusher Funnel Loop
Systematically identifies and counters customer hesitation points across a multi-stage landing page.
The Feature-Benefit Stress Test
Challenges every property feature with the 'So What?' question to find real selling points.
The Referral Loop Leakage Sweep
Identifies gaps in the specialist referral chain to ensure no patient falls through the cracks.
The Data-Backed Credibility Builder
Turns anecdotal claims into authoritative, data-supported arguments.
The Zero-Click Content Loop
Transforms long-form blog posts into platform-native threads that satisfy the algorithm.
The Lab Result Translator Loop
Converts complex lab values into easy-to-understand summaries for the patient portal.
The Multi-Threader Architect
Maps out a strategy to engage multiple stakeholders within a single account.
The Handoff Harmonizer
Converts complex patient histories into the structured I-PASS format for safe shift changes.
The Referral Loop Closer
Automates the identification and thanking of referral sources to keep the pipeline full.
Frequently asked
Yes. They work in Google AI Studio, the Gemini consumer app, Vertex AI, and the Gemini API.
Many of our prompts are explicitly designed for 200k+ token inputs - look for the 'long-context' tag.