AI on Frytea

终极指南：在 Linux 裸机服务器上快速部署 Moltbot (原 Clawbot) 并集成飞书

Thu, 29 Jan 2026 12:00:15 +0000

引言

2026 年初，一个名为 Moltbot（前身为 Clawbot）的开源 AI 代理（Agent）框架席卷了开发者社区。该框架允许用户将强大的 AI 模型（例如 OpenAI 的 GPT 系列、Anthropic 的 Claude 等）与 WhatsApp、Telegram、Discord 等日常通讯工具集成，从而通过聊天即可操控电脑、执行任务、获取信息。特斯拉前 AI 主管 Andrej Karpathy 的推荐更是使其迅速走红，其 GitHub 项目在短时间内获得了超过 60,000 个星标。

AI 生图精品提示词｜第一期

Sat, 06 Dec 2025 06:47:00 +0000

若没有特别说明，默认使用 AiLoft 提供的 Nano Banana Pro 模型生成。

本期封面图：

3D Q 版迷你概念商店

Refs: https://x.com/dotey/status/1995190286775881780

说明：

🍌 nano banana prompt
3D chibi-style miniature concept store of {Brand Name}

提示词：

3D chibi-style miniature concept store of {Brand Name}, creatively designed with an exterior inspired by the brand's most iconic product or packaging (such as a giant {brand's core product, e.g., chicken bucket/hamburger/donut/roast duck}). The store features two floors with large glass windows clearly showcasing the cozy and finely decorated interior: {brand's primary color}-themed decor, warm lighting, and busy staff dressed in outfits matching the brand. Adorable tiny figures stroll or sit along the street, surrounded by benches, street lamps, and potted plants, creating a charming urban scene. Rendered in a miniature cityscape style using Cinema 4D, with a blind-box toy aesthetic, rich in details and realism, and bathed in soft lighting that evokes a relaxing afternoon atmosphere. --ar 2:3

Brand name: Starbucks

效果图： Starbucks

整理了一个 AI 提示词库

Sun, 07 Sep 2025 17:00:00 +0000

大模型的能力越来越强，但是如何发挥大模型真正的实力？

这是我一直在思考的问题，在平常的使用过程中很容易发现，提示词的使用技巧对于生成结果的质量至关重要。多看看优秀的提示词，还能启发我们利用大模型的更多用法。

让 LLM 看到真实世界的 Playwright MCP

Sun, 07 Sep 2025 14:22:41 +0000

Playwright MCP 是一个模型上下文协议（MCP）服务器，使用 Playwright 提供浏览器自动化功能。该服务器使 LLM 能够通过结构化的可访问性快照与网页交互，从而绕过对屏幕截图或视觉调整模型的需求。

Claude Code Subagents 快速开始

Tue, 29 Jul 2025 01:26:08 +0000

在Claude Code中的AI子代理，可用于特定任务的工作流程和改进的上下文管理。

Claude Code中的自定义子代理是专门的AI助手，可以被调用来处理特定类型的任务。它们通过提供具有自定义系统提示、工具和独立上下文窗口的特定任务配置，实现更高效的问题解决。

Lyra - AI Prompt Optimization Specialist

Sun, 20 Jul 2025 10:09:00 +0000

AI 提示优化专家 - Lyra, 一个很好的提示词。

TL;DR

You are Lyra, a master-level AI prompt optimization specialist. Your mission: transform any user input into
precision-crafted prompts that unlock Al's full potential across all platforms.

## THE 4-D METHODOLOGY

### 1. DECONSTRUCT
- Extract core intent, key entities, and context
- Identify output requirements and constraints
- Map what's provided vs. what's missing

### 2. DIAGNOSE
- Audit for clarity gaps and ambiguity
- Check specificity and completeness
- Assess structure and complexity needs

### 3. DEVELOP
- Select optimal techniques based on request type:
- **Creative** → Multi-perspective + tone emphasis
- **Technical** → Constraint-based + precision focus
- **Educational** → Few-shot examples + clear structure
- **Complex** → Chain-of-thought + systematic frameworks
- Assign appropriate Al role/expertise
- Enhance context and implement logical structure

### 4. DELIVER
- Construct optimized prompt
- Format based on complexity
- Provide implementation guidance

## OPTIMIZATION TECHNIQUES

**Foundation:** Role assignment, context layering, output specs, task decomposition

**Advanced:** Chain-of-thought, few-shot learning, multi-perspective analysis, constraint optimization

**Platform Notes:**
- **ChatGPT/GPT-4:** Structured sections, conversation starters
- **Claude:** Longer context, reasoning frameworks
- **Gemini:** Creative tasks, comparative analysis
- **Others:** Apply universal best practices

## OPERATING MODES

**DETAIL MODE:**
- Gather context with smart defaults
- Ask 2-3 targeted clarifying questions
- Provide comprehensive optimization

**BASIC MODE:**
- Quick fix primary issues
- Apply core techniques only
- Deliver ready-to-use prompt

## RESPONSE FORMATS

**Simple Requests:**
---
**Your Optimized Prompt:**
[Improved prompt]
**What Changed:** [Key improvements]
---

**Complex Requests:**
---
**Your Optimized Prompt:**
[Improved prompt]
**Key Improvements:**
• [Primary changes and benefits]
**Techniques Applied:** [Brief mention]
**Pro Tip:** [Usage guidance]
---

## WELCOME MESSAGE (REQUIRED)

When activated, display EXACTLY:

"Hello! I'm Lyra, your Al prompt optimizer. I transform vague requests into precise, effective prompts that
deliver better results.

**What I need to know:**
- **Target AI:** ChatGPT, Claude, Gemini, or Other
- **Prompt Style:** DETAIL (I'll ask clarifying questions first) or BASIC (quick optimization)

**Examples:**
- "DETAIL using ChatGPT - Write me a marketing email"
- "BASIC using Claude - Help with my resume"

Just share your rough prompt and I'll handle the optimization!"

## PROCESSING FLOW

1. Auto-detect complexity:
   - Simple tasks → BASIC mode
   - Complex/professional → DETAIL mode
2. Inform user with override option
3. Execute chosen mode protocol (see below)
4. Deliver optimized prompt

**Memory Note:** Do not save any information from optimization sessions to memory.

References

https://gist.github.com/xthezealot/c873effd9e74225ef3fcfbb9c3a341da

判断GPT是否降智的几个问题

Thu, 05 Jun 2025 09:19:00 +0000

判别方法

以下问题，问3-4次

要多问几次，有的第一次不会降，其实问一次就被标记了，新开两三个对话再问一下

方法一：小数点数字比大小

6.9 和 6.11 哪一个大？

降智回答：6.11大
正确回答：也会出错，但会分析纠正，得出6.9更大

OpenManus 使用记录

Mon, 21 Apr 2025 08:35:45 +0000

安装运行

# 安装 uv（一个快速的 Python 包管理器）：
$ curl -LsSf https://astral.sh/uv/install.sh | sh
# 克隆仓库：
$ git clone https://github.com/mannaandpoem/OpenManus.git
$ cd OpenManus
# 创建并激活虚拟环境：
$ uv venv --python 3.12
$ source .venv/bin/activate  # Unix/macOS 系统
# Windows 系统使用：
# .venv\Scripts\activate
# 安装依赖：
$ uv pip install -r requirements.txt
# 浏览器自动化工具（可选）
$ playwright install
# 在 config 目录创建 config.toml 文件（可从示例复制）：
$ cp config/config.example.toml config/config.toml
$ vim config/config.toml
# 一行命令运行 OpenManus：
$ python main.py

实际效果

开始执行：

绘图模型效果对比之城市气象

Mon, 21 Apr 2025 08:34:00 +0000

Promot

A highly detailed and photorealistic image of Akmenė, Pemagatshel, Lithuania, during a cloudy day with a temperature feel of 7.0°C. The scene captures the historic Church of St. John the Baptist, its intricate brickwork and architectural features glistening from recent rain. The wet pavement reflects the overcast sky, which is 76.0% cloud covered, creating a soft, diffused light that highlights the textures of the buildings and surrounding lush greenery. The foreground includes puddles forming on the cobblestone streets, while the background features dense, misty forests framing the town. The composition employs the rule of thirds, with the church positioned to the right, leading lines from the street guiding the viewer’s eye through the scene. The lighting enhances the mood, maintaining clarity and brightness, with details visible in the shadows. The overall atmosphere is tranquil and inviting, embodying the rich cultural heritage and natural beauty of Akmenė.

Refs: Akmenė, Lithuania Weather Art

LLM 聚合 API 价格对比

Fri, 14 Mar 2025 17:26:01 +0000

List

gpt-4
gpt-4o
claude-3-7-sonnet-20250219
`claude-3-7-sonn

单位：Inout/Output /M

Model	`deepseek-r1`	`claude-3-7-sonnet`	`claude-3-5-sonnet`
UniAPI	`$0.2871/$1.1484`	`$2.376/$11.88`	`$2.376/$11.88`
GPTAPI	`¥0.07/¥0.14`	`¥5.25/¥26.25`	`¥5.25/¥26.25`
OpenRouter	`$5/$7`	`$3/$15`	`$3/$15`
AiHubMix	`$0.62/$2.48`	`$3.3/$16.5`
V3 API	`$1.8/$7.2`	`$7.4/$37`

Refereneces

全球艺术天气网站

Sun, 02 Feb 2025 15:08:00 +0000

介绍一个每日 AI 艺术天气网站，能够通过人工智能生成的艺术作品，以独特的视角展现日常气象现象。

入口

TodayAiWeather

部分城市入口：

收录情况

目前收录了 78 个城市，未来还会根据需要添加更多，有需要可以在本文留言。

使用 GCOP AI 助力提升 git commit 效率

Sat, 04 Jan 2025 01:54:12 +0000

最近在阮老师博客看到一款这个 GCOP，一个基于 LLM 的 Git 提交消息生成和 Git 工作流程优化工具，试用了一下确实不错。接入 LLM 即可自动分析 git 修改生成 commit 消息，比我之前做的 git-commit 生成器好用多了。

【Ai Agent】智能体实验之城市天气画报

Fri, 27 Dec 2024 10:19:00 +0000

计划进行一系列 AI Agent 实验，欢迎交流。

DSL

app:
  description: ''
  icon: 🤖
  icon_background: '#FFEAD5'
  mode: agent-chat
  name: 城市天气画报
  use_icon_as_answer_icon: false
kind: app
model_config:
  agent_mode:
    enabled: true
    max_iteration: 5
    prompt: null
    strategy: function_call
    tools:
    - enabled: true
      provider_id: gaode
      provider_name: gaode
      provider_type: builtin
      tool_label: 天气预报
      tool_name: gaode_weather
      tool_parameters:
        city: ''
    - enabled: true
      provider_id: dalle
      provider_name: dalle
      provider_type: builtin
      tool_label: DALL-E 3 绘画
      tool_name: dalle3
      tool_parameters:
        n: ''
        prompt: ''
        quality: ''
        size: ''
        style: ''
  annotation_reply:
    enabled: false
  chat_prompt_config: {}
  completion_prompt_config: {}
  dataset_configs:
    datasets:
      datasets: []
    reranking_enable: true
    retrieval_model: multiple
    top_k: 4
  dataset_query_variable: ''
  external_data_tools: []
  file_upload:
    allowed_file_extensions:
    - .JPG
    - .JPEG
    - .PNG
    - .GIF
    - .WEBP
    - .SVG
    - .MP4
    - .MOV
    - .MPEG
    - .MPGA
    allowed_file_types: []
    allowed_file_upload_methods:
    - remote_url
    - local_file
    enabled: false
    image:
      detail: high
      enabled: false
      number_limits: 3
      transfer_methods:
      - remote_url
      - local_file
    number_limits: 3
  model:
    completion_params:
      stop: []
    mode: chat
    name: gpt-4o-mini
    provider: openai
  more_like_this:
    enabled: false
  opening_statement: 回复 1 开始创作今日城市天气画报。
  pre_prompt: '查找城市 {{city}} 今天的天气情况，生成一段简单的话描述这个场景，提供给 dalle 绘制一幅图片。

    '
  prompt_type: simple
  retriever_resource:
    enabled: true
  sensitive_word_avoidance:
    configs: []
    enabled: false
    type: ''
  speech_to_text:
    enabled: false
  suggested_questions: []
  suggested_questions_after_answer:
    enabled: false
  text_to_speech:
    enabled: false
    language: ''
    voice: ''
  user_input_form:
  - text-input:
      default: ''
      label: 城市
      max_length: 48
      required: true
      variable: city
version: 0.1.5

效果展示

界面：

自托管 AIGC UI APP 汇总

Fri, 20 Dec 2024 04:16:00 +0000

该清单搜集人类创造的最先进最好用的开源 AIGC APP 方案，包括但不限于对话、识图、生图、TTS、知识库、多模态、工作流编排，主要搜集支持自托管，可容器化部署的方案。

AI Prompt 整理

Thu, 12 Dec 2024 10:50:00 +0000

prompt 原则

1、写出清晰而具体的指示，清晰不是指短，而是尽量详细，可采用分隔符避免自己的要求和内容的混淆，分隔符形式不限：“””, < >, {} 2、给模型思考的时间。如果答案不对，可以尝试不断变化提问方式，直到得出想要结果。可以采用简化或分解问题的方式。