Back to Details

gemini-text

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Gemini Text Generation

Gemini文本生成

Generate content using Google's Gemini API through executable scripts with advanced capabilities including system instructions, thinking mode, JSON output, and Google Search grounding.
通过可执行脚本使用Google的Gemini API生成内容，具备系统指令、思维模式、JSON输出和Google Search grounding等高级功能。

When to Use This Skill

何时使用该技能

Use this skill when you need to:
  • Generate any type of text content (blogs, emails, code, stories)
  • Process images with text descriptions or analysis
  • Perform complex reasoning requiring step-by-step thinking
  • Get structured JSON outputs for data processing
  • Access real-time information via Google Search
  • Apply specific personas or behavior patterns
  • Combine text generation with other Gemini skills (images, TTS, embeddings)
当你需要以下功能时使用本技能：
  • 生成任意类型的文本内容（博客、邮件、代码、故事）
  • 结合文本描述或分析处理图片
  • 执行需要逐步思考的复杂推理
  • 获取用于数据处理的结构化JSON输出
  • 通过Google Search获取实时信息
  • 应用特定角色或行为模式
  • 将文本生成与其他Gemini技能（图片、TTS、嵌入）结合使用

Available Scripts

可用脚本

scripts/generate.py

scripts/generate.py

Purpose: Full-featured text generation with all Gemini capabilities
When to use:
  • Any text generation task
  • Multimodal prompts (text + image)
  • Complex reasoning requiring thinking mode
  • Structured JSON output requirements
  • Real-time information needs (grounding)
  • Custom system instructions/personas
Key parameters:
ParameterDescriptionExample
prompt
Text prompt (required)
"Explain quantum computing"
--model
, 
-m
Model to use
gemini-3-flash-preview
--system
, 
-s
System instruction
"You are a helpful assistant"
--thinking
, 
-t
Enable thinking modeFlag
--json
, 
-j
Force JSON outputFlag
--grounding
, 
-g
Enable Google SearchFlag
--image
, 
-i
Image for multimodal
photo.png
--temperature
Sampling 0.0-2.0
0.7
for creative
--max-tokens
Output limit
1000
Output: Generated text string, optionally with grounding sources
用途：具备所有Gemini功能的全特性文本生成工具
适用场景：
  • 任意文本生成任务
  • 多模态提示词（文本+图片）
  • 需要思维模式的复杂推理
  • 结构化JSON输出需求
  • 实时信息获取需求（grounding）
  • 自定义系统指令/角色
关键参数：
参数描述示例
prompt
文本提示词（必填）
"Explain quantum computing"
--model
, 
-m
使用的模型
gemini-3-flash-preview
--system
, 
-s
系统指令
"You are a helpful assistant"
--thinking
, 
-t
启用思维模式标志位
--json
, 
-j
强制输出JSON标志位
--grounding
, 
-g
启用Google Search标志位
--image
, 
-i
用于多模态的图片
photo.png
--temperature
采样温度（0.0-2.0）
0.7
（适合创意内容）
--max-tokens
输出长度限制
1000
输出：生成的文本字符串，可附带grounding来源

Workflows

工作流

Workflow 1: Basic Text Generation

工作流1：基础文本生成

bash
python scripts/generate.py "Explain quantum computing in simple terms"
  • Best for: Simple content creation, explanations, summaries
  • Model: 
    gemini-3-flash-preview
    (default, fast)
bash
python scripts/generate.py "Explain quantum computing in simple terms"
  • 最佳适用：简单内容创作、解释、摘要
  • 默认模型：
    gemini-3-flash-preview
    （速度快）

Workflow 2: With System Instruction (Persona)

工作流2：带系统指令（角色设定）

bash
python scripts/generate.py "How do I read a file in Python?" --system "You are a helpful coding assistant"
  • Best for: Domain-specific tasks, expert personas, consistent tone
  • Use when: You need specific behavioral constraints
bash
python scripts/generate.py "How do I read a file in Python?" --system "You are a helpful coding assistant"
  • 最佳适用：特定领域任务、专家角色、一致语气
  • 适用场景：需要特定行为约束时

Workflow 3: Complex Reasoning (Thinking Mode)

工作流3：复杂推理（思维模式）

bash
python scripts/generate.py "Analyze the ethical implications of AI in healthcare" --thinking
  • Best for: Complex analysis, step-by-step reasoning, multi-step problems
  • Use when: Task requires careful consideration and logical progression
bash
python scripts/generate.py "Analyze the ethical implications of AI in healthcare" --thinking
  • 最佳适用：复杂分析、逐步推理、多步骤问题
  • 适用场景：任务需要细致考量和逻辑递进时

Workflow 4: Structured JSON Output

工作流4：结构化JSON输出

bash
python scripts/generate.py "Generate a user profile object with name, email, and preferences" --json
  • Best for: Data extraction, structured data generation, API responses
  • Output: Valid JSON ready for parsing
  • Note: Prompt must clearly request JSON structure
bash
python scripts/generate.py "Generate a user profile object with name, email, and preferences" --json
  • 最佳适用：数据提取、结构化数据生成、API响应
  • 输出：可直接解析的有效JSON
  • 注意：提示词必须明确要求JSON结构

Workflow 5: Real-Time Information (Grounding)

工作流5：实时信息获取（Grounding）

bash
python scripts/generate.py "Who won the latest Super Bowl?" --grounding
  • Best for: Current events, news, factual information after training cutoff
  • Output: Response + grounding sources with citations
  • Use when: Accuracy of current information is critical
bash
python scripts/generate.py "Who won the latest Super Bowl?" --grounding
  • 最佳适用：当前事件、新闻、训练 cutoff 后的事实信息
  • 输出：响应内容 + 带引用的grounding来源
  • 适用场景：当前信息的准确性至关重要时

Workflow 6: Multimodal (Image Analysis)

工作流6：多模态（图片分析）

bash
python scripts/generate.py "Describe what's in this image in detail" --image photo.png
  • Best for: Image captioning, visual analysis, image-based Q&A
  • Requires: Image file in PNG or JPEG format
  • Combines well with: gemini-files for file upload
bash
python scripts/generate.py "Describe what's in this image in detail" --image photo.png
  • 最佳适用：图片标题生成、视觉分析、基于图片的问答
  • 要求：PNG或JPEG格式的图片文件
  • 最佳搭配：gemini-files技能用于文件上传

Workflow 7: Content Creation Pipeline (Batch + Text + TTS)

工作流7：内容创作流水线（批量+文本+TTS）

bash
undefined
bash
undefined

1. Create batch requests (gemini-batch skill)

1. 创建批量请求（gemini-batch技能）

2. Generate content

2. 生成内容

python scripts/generate.py "Create a 500-word blog post about sustainable energy"
python scripts/generate.py "Create a 500-word blog post about sustainable energy"

3. Convert to audio (gemini-tts skill)

3. 转换为音频（gemini-tts技能）

- Best for: High-volume content production, podcasts, audiobooks
- 最佳适用：高容量内容生产、播客、有声书

Parameters Reference

参数参考

Model Selection

模型选择

ModelSpeedIntelligenceContextBest For
gemini-3-flash-preview
FastHigh1MGeneral use, agentic tasks (default)
gemini-3-pro-preview
MediumHighest1MComplex reasoning, research
gemini-2.5-flash
FastMedium1MStable, reliable generation
gemini-2.5-pro
SlowHigh1MCode, math, STEM tasks
模型速度智能程度上下文长度最佳适用场景
gemini-3-flash-preview
快高1M通用场景、智能体任务（默认）
gemini-3-pro-preview
中等最高1M复杂推理、研究
gemini-2.5-flash
快中等1M稳定、可靠的生成
gemini-2.5-pro
慢高1M代码、数学、STEM任务

Temperature Settings

温度设置

ValueCreativityBest For
0.0-0.3LowCode, facts, formal writing
0.4-0.7MediumBalanced output
0.8-1.0HighCreative writing, brainstorming
1.0-2.0Very HighHighly creative, varied outputs
数值创意程度最佳适用场景
0.0-0.3低代码、事实、正式写作
0.4-0.7中等平衡型输出
0.8-1.0高创意写作、头脑风暴
1.0-2.0极高高度创意、多样化输出

Thinking Budget

思维预算

ValueDescription
0Disabled (default behavior)
512-1024Standard reasoning
2048+Deep analysis (slower, more tokens)
数值描述
0禁用（默认行为）
512-1024标准推理
2048+深度分析（速度慢、消耗更多tokens）

Output Interpretation

输出解读

Standard Text Output

标准文本输出

  • Plain text response ready for use
  • Check for truncation if max-tokens was set
  • May include markdown formatting
  • 可直接使用的纯文本响应
  • 如果设置了max-tokens，检查是否被截断
  • 可能包含Markdown格式

JSON Output

JSON输出

  • Valid JSON object (use 
    --json
    flag)
  • Parse with: 
    import json; data = json.loads(output)
  • Verify structure matches your requirements
  • Handle potential parsing errors
  • 有效的JSON对象（使用
    --json
    标志位）
  • 解析方式：
    import json; data = json.loads(output)
  • 验证输出结构是否符合需求
  • 处理潜在的解析错误

Grounded Response

Grounded响应

When 
--grounding
is used, the script prints:
  1. Main response text
  2. "--- Grounding Sources ---" section
  3. List of sources with titles and URLs
当使用
--grounding
时，脚本会输出：
  1. 主要响应文本
  2. "--- Grounding Sources ---" 部分
  3. 包含标题和URL的来源列表

Thinking Mode Output

思维模式输出

  • May include reasoning steps before final answer
  • Longer response times due to thinking process
  • Better for tasks requiring careful analysis
  • 最终答案前可能包含推理步骤
  • 由于思考过程，响应时间更长
  • 更适合需要细致分析的任务

Common Issues

常见问题

"google-genai not installed"

"google-genai not installed"

bash
pip install google-genai
bash
pip install google-genai

"API key not set"

"API key not set"

Set environment variable:
bash
export GOOGLE_API_KEY="your-key-here"
设置环境变量：
bash
export GOOGLE_API_KEY="your-key-here"

or

or

export GEMINI_API_KEY="your-key-here"
undefined
export GEMINI_API_KEY="your-key-here"
undefined

"Model not available"

"Model not available"

  • Check model name spelling
  • Verify API access for selected model
  • Try 
    gemini-3-flash-preview
    (most available)
  • 检查模型名称拼写
  • 验证所选模型的API访问权限
  • 尝试使用
    gemini-3-flash-preview
    （可用性最高）

JSON parse errors

JSON解析错误

  • Ensure prompt explicitly requests JSON structure
  • Check output for JSON formatting
  • Consider using system instruction: "You always respond with valid JSON"
  • 确保提示词明确要求JSON结构
  • 检查输出的JSON格式
  • 考虑使用系统指令："You always respond with valid JSON"

Image file not found

图片文件未找到

  • Verify image path is correct
  • Use absolute paths if relative paths fail
  • Supported formats: PNG, JPEG
  • 验证图片路径是否正确
  • 如果相对路径失败，使用绝对路径
  • 支持的格式：PNG、JPEG

Response truncated

响应被截断

  • Increase 
    --max-tokens
    value
  • Break task into smaller requests
  • Use pro models with higher token limits
  • 提高
    --max-tokens
    的值
  • 将任务拆分为更小的请求
  • 使用支持更高token限制的pro模型

Best Practices

最佳实践

Performance Optimization

性能优化

  • Use flash models for speed, pro for quality
  • Lower temperature (0.0-0.3) for deterministic outputs
  • Set appropriate max-tokens to control costs
  • Use thinking mode only for complex tasks
  • 追求速度用flash模型，追求质量用pro模型
  • 降低温度（0.0-0.3）以获得确定性输出
  • 设置合适的max-tokens以控制成本
  • 仅在复杂任务中使用思维模式

Prompt Engineering

提示词工程

  • Be specific and clear in your prompts
  • Use system instructions for consistent behavior
  • Include examples in prompts for better results
  • For JSON: specify exact structure in prompt
  • 提示词要具体清晰
  • 使用系统指令确保行为一致
  • 提示词中包含示例以获得更好的结果
  • 生成JSON时：在提示词中指定确切的结构

Error Handling

错误处理

  • Wrap script calls in try-except blocks
  • Validate JSON output before parsing
  • Handle network timeouts with retries
  • Check API quota limits for batch operations
  • 将脚本调用包裹在try-except块中
  • 解析前验证JSON输出
  • 处理网络超时，实现重试机制
  • 批量操作时检查API配额限制

Cost Management

成本管理

  • Use flash models when possible (lower cost)
  • Limit max-tokens for simple queries
  • Cache results for repeated queries
  • Use batch API for high-volume tasks
  • 尽可能使用flash模型（成本更低）
  • 简单查询限制max-tokens
  • 缓存重复查询的结果
  • 高容量任务使用批量API

Related Skills

相关技能

  • gemini-image: Generate images from text
  • gemini-tts: Convert text to speech
  • gemini-embeddings: Create vector embeddings for semantic search
  • gemini-files: Upload files for multimodal processing
  • gemini-batch: Process multiple requests efficiently
  • gemini-image：从文本生成图片
  • gemini-tts：将文本转换为语音
  • gemini-embeddings：创建用于语义搜索的向量嵌入
  • gemini-files：上传文件用于多模态处理
  • gemini-batch：高效处理多个请求

Quick Reference

快速参考

bash
undefined
bash
undefined

Basic

Basic

python scripts/generate.py "Your prompt"
python scripts/generate.py "Your prompt"

Persona

Persona

python scripts/generate.py "Prompt" --system "You are X"
python scripts/generate.py "Prompt" --system "You are X"

Thinking

Thinking

python scripts/generate.py "Complex task" --thinking
python scripts/generate.py "Complex task" --thinking

JSON

JSON

python scripts/generate.py "Generate JSON" --json
python scripts/generate.py "Generate JSON" --json

Search

Search

python scripts/generate.py "Current event" --grounding
python scripts/generate.py "Current event" --grounding

Multimodal

Multimodal

python scripts/generate.py "Describe this" --image photo.png
undefined
python scripts/generate.py "Describe this" --image photo.png
undefined

Reference

参考资料