# Griptape — Modular Python AI Agent Framework

> Build AI agents with composable structures, tools, and memory. Off-Prompt data processing for secure enterprise use. 2.5K+ stars.

## Install

Save the content below to `.claude/skills/` or append to your `CLAUDE.md`:

# Griptape — Modular Python AI Agent Framework

## Quick Use

```bash
pip install griptape[all]
```

```python
from griptape.structures import Agent
from griptape.tools import WebScraper, Calculator, FileManager

# Create an agent with tools
agent = Agent(
    tools=[
        WebScraper(),
        Calculator(),
        FileManager(),
    ]
)

agent.run("Scrape the Hacker News front page, calculate the average score of the top 5 posts, and save the result to results.txt")
```

Set your API key:

```bash
export OPENAI_API_KEY="sk-..."
# Or for Claude:
export ANTHROPIC_API_KEY="sk-ant-..."
```

---

## Intro

Griptape is a modular Python framework with 2,500+ GitHub stars for building AI agents, pipelines, and workflows. Its key innovation is **Off-Prompt** — a pattern where sensitive data is processed outside the LLM context window, enabling secure enterprise deployments. Griptape provides three composable structures (Agent, Pipeline, Workflow) that can be mixed with pluggable tools, memory modules, and any LLM backend. Backed by a company offering Griptape Cloud for managed deployment, it bridges the gap between experimental AI agents and production enterprise systems.

Works with: OpenAI, Anthropic Claude, Google Gemini, AWS Bedrock, Azure OpenAI, Cohere, HuggingFace, Ollama. Best for enterprise teams building AI agents that handle sensitive data securely. Setup time: under 3 minutes.

---

## Griptape Structures & Patterns

### Three Core Structures

```python
from griptape.structures import Agent, Pipeline, Workflow

# 1. Agent — Single task with tools
agent = Agent(tools=[WebScraper(), Calculator()])
agent.run("What's 15% tip on a $85 dinner?")

# 2. Pipeline — Sequential multi-step tasks
pipeline = Pipeline()
pipeline.add_tasks(research_task, analyze_task, report_task)
pipeline.run("Analyze Q4 sales data")

# 3. Workflow — Parallel + sequential DAG
workflow = Workflow()
workflow.add_tasks(task_a, task_b, task_c)  # a,b run in parallel
task_c.add_parents([task_a, task_b])         # c waits for both
workflow.run()
```

### Off-Prompt Processing

Griptape's signature feature — keep sensitive data out of the LLM:

```python
from griptape.structures import Agent
from griptape.drivers import LocalVectorStoreDriver
from griptape.engines import VectorQueryEngine

# Data stays in local vector store, never sent to LLM
vector_store = LocalVectorStoreDriver()
query_engine = VectorQueryEngine(vector_store_driver=vector_store)

agent = Agent(
    tools=[
        # Tool processes data locally, sends only summaries to LLM
        VectorStoreTool(query_engine=query_engine, off_prompt=True)
    ]
)
agent.run("What are the key risks in our financial report?")
# LLM sees: "Based on the retrieved context: [summary]"
# LLM never sees: raw financial data
```

### Tool System

Pre-built tools for common agent tasks:

| Tool | Capability |
|------|-----------|
| **WebScraper** | Fetch and parse web pages |
| **Calculator** | Mathematical operations |
| **FileManager** | Read/write local files |
| **SqlClient** | Query SQL databases |
| **VectorStoreTool** | Semantic search over documents |
| **AwsS3Client** | Read/write S3 objects |
| **GoogleCalendar** | Manage calendar events |
| **EmailClient** | Send and read emails |
| **RestApiClient** | Call any REST API |

Build custom tools:

```python
from griptape.tools import BaseTool
from griptape.artifacts import TextArtifact
from griptape.utils.decorators import activity
from schema import Schema, Literal

class StockPriceTool(BaseTool):
    @activity(
        config={"description": "Get current stock price"},
        schema=Schema({Literal("symbol"): str})
    )
    def get_price(self, params):
        symbol = params["values"]["symbol"]
        price = fetch_stock_price(symbol)
        return TextArtifact(f"{symbol}: ${price}")
```

### Memory & Conversation

```python
from griptape.structures import Agent
from griptape.memory.structure import ConversationMemory
from griptape.memory.task.storage import LocalDiskTaskMemoryDriver

# Persistent conversation memory
agent = Agent(
    conversation_memory=ConversationMemory(
        driver=LocalDiskTaskMemoryDriver(file_path="memory.json")
    )
)

agent.run("My name is Alice and I work on ML pipelines")
# Later...
agent.run("What do you know about me?")
# -> "You're Alice, you work on ML pipelines"
```

### Multi-LLM Support

```python
from griptape.drivers import AnthropicPromptDriver, OpenAiChatPromptDriver

# Use Claude for reasoning
agent = Agent(
    prompt_driver=AnthropicPromptDriver(model="claude-sonnet-4-6")
)

# Or OpenAI
agent = Agent(
    prompt_driver=OpenAiChatPromptDriver(model="gpt-4o")
)

# Or local via Ollama
from griptape.drivers import OllamaPromptDriver
agent = Agent(
    prompt_driver=OllamaPromptDriver(model="llama3")
)
```

---

## FAQ

**Q: What is Griptape?**
A: Griptape is a modular Python framework with 2,500+ GitHub stars for building AI agents with composable structures (Agent, Pipeline, Workflow), pluggable tools, and Off-Prompt data processing for secure enterprise use.

**Q: What makes Griptape different from LangChain or CrewAI?**
A: Griptape's key differentiator is Off-Prompt processing — sensitive data is processed locally and only summaries reach the LLM. It also has a cleaner modular architecture with three distinct structures, versus LangChain's sprawling API surface. CrewAI focuses on multi-agent roles; Griptape focuses on secure, composable pipelines.

**Q: Is Griptape free?**
A: Yes, the framework is open-source under Apache-2.0. Griptape Cloud (managed hosting) is a paid service with a free tier.

---

## Source & Thanks

> Created by [Griptape](https://github.com/griptape-ai). Licensed under Apache-2.0.
>
> [griptape](https://github.com/griptape-ai/griptape) — ⭐ 2,500+

Thanks to the Griptape team for bringing enterprise-grade security patterns to AI agent development.

---

<!-- ZH -->

## 快速使用

```bash
pip install griptape[all]
```

```python
from griptape.structures import Agent
from griptape.tools import WebScraper, Calculator, FileManager

agent = Agent(tools=[WebScraper(), Calculator(), FileManager()])
agent.run("抓取 Hacker News 首页，计算前5条的平均分数，保存到 results.txt")
```

---

## 简介

Griptape 是一个拥有 2,500+ GitHub stars 的模块化 Python AI 代理框架。其核心创新是 **Off-Prompt** 模式 — 敏感数据在 LLM 上下文窗口外处理，确保企业级数据安全。提供三种可组合结构（Agent、Pipeline、Workflow），搭配可插拔的工具、记忆模块和任意 LLM 后端。

适用于：OpenAI、Anthropic Claude、Gemini、AWS Bedrock、Ollama。适合需要安全处理敏感数据的企业 AI 代理开发团队。

---

## 核心特性

### 三种结构
- **Agent** — 单任务代理，配备工具
- **Pipeline** — 顺序多步任务
- **Workflow** — 并行+顺序 DAG 执行

### Off-Prompt 数据处理
敏感数据在本地处理，只有摘要发送给 LLM，确保数据安全。

### 丰富的工具生态
Web 抓取、数据库查询、文件管理、S3、邮件、日历等预置工具。

### 多 LLM 支持
OpenAI、Anthropic、Gemini、Bedrock、Ollama 等主流提供商。

---

## 来源与感谢

> Created by [Griptape](https://github.com/griptape-ai). Licensed under Apache-2.0.
>
> [griptape](https://github.com/griptape-ai/griptape) — ⭐ 2,500+


---
Source: https://tokrepo.com/en/workflows/43de1956-a080-47df-bf44-087257ebf252
Author: TokRepo精选