How do I write a good Gemini system prompt?

Gemini responds well to clear, structured prompts. Use separate sections for the AI's role, the task, context, and the desired output format. Markdown headers like # Role and # Task work reliably.

What is the Gemini 1.5 Pro context window?

Gemini 1.5 Pro supports a 2,000,000-token context window — one of the largest available. This allows extremely long system prompts with many examples.

Does Gemini support system prompts in the API?

Yes. The Gemini API supports a system instruction field that accepts plain text or structured content. This prompt builder assembles content suitable for that field.

Gemini Prompt Builder

Build structured prompts for Google Gemini AI models.

Role / Persona

Task / Instruction

Context / Background

Constraints / Rules

Output Format

Examples

Assembled Prompt

0 tokens · 0 chars

Fill in the fields on the left to build your prompt...

Format: Markdown headers

Related Tools

PBDAI Prompt BuilderNEW

Build structured AI prompts with role, task, context, and output format fields.

CLDClaude Prompt BuilderNEW

Build structured system prompts for Anthropic Claude using XML tags.

GEMGemini Token CounterNEW

Count tokens for Gemini 1.5 Pro, Flash, and Gemini 1.0 Pro models.

GEMGemini API Cost CalculatorNEW

Calculate Google Gemini API costs for 1.5 Pro, Flash, and 1.0 Pro.

OPTAI Prompt OptimizerNEW

Analyze and improve AI prompts with rule-based suggestions.

Learn More

guide:prompt engineering basics

FAQ

How do I write a good Gemini system prompt?: Gemini responds well to clear, structured prompts. Use separate sections for the AI's role, the task, context, and the desired output format. Markdown headers like # Role and # Task work reliably.
What is the Gemini 1.5 Pro context window?: Gemini 1.5 Pro supports a 2,000,000-token context window — one of the largest available. This allows extremely long system prompts with many examples.
Does Gemini support system prompts in the API?: Yes. The Gemini API supports a system instruction field that accepts plain text or structured content. This prompt builder assembles content suitable for that field.

Build well-structured prompts optimized for Google Gemini models. Assembled using markdown headers — a clear format that works well with Gemini 1.5 Pro and Gemini 1.5 Flash. Fill in role, task, context, constraints, output format, and examples.