# AgentKit Samples — Hello-World Agent (uv + SSE)

> AgentKit Samples is a code workshop repo; the Hello World example uses `uv` to run a chat agent with short-term memory and an SSE client for testing.

## Install

Copy the content below into your project:

## Quick Use

```bash
git clone https://github.com/bytedance/agentkit-samples.git
cd agentkit-samples/02-use-cases/beginner/hello_world

curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv --python 3.12
uv sync

uv run agent.py
```

## Intro

AgentKit Samples is a code workshop repo; the Hello World example uses `uv` to run a chat agent with short-term memory and an SSE client for testing.

- **Best for:** Developers who want runnable AgentKit examples (memory, tools, MCP) rather than marketing diagrams
- **Works with:** Python 3.10+; `uv`; Volcengine credentials for model access in examples (per sample README)
- **Setup time:** 15–30 minutes

## Practical Notes

- GitHub: 330 stars · 60 forks; pushed 2026-05-12 (verified via GitHub API).
- Hello World sample README uses `uv venv` + `uv sync`, then `uv run agent.py` to start the agent service.
- Sample includes an SSE client (`client.py`) and environment variables for model selection/credentials (per README).

## Main

A good way to learn from this repo:

1. Pick one “beginner” use case (like `hello_world`) and get it running end-to-end first.
2. Read the example’s README carefully—many samples include provider-specific credentials and deployment steps.
3. Once it runs locally, use it as a template:
   - swap the memory backend
   - add one tool integration
   - then add deployment config
4. Keep a notebook of the exact commands you used (`uv venv`, `uv sync`, `uv run …`) so your teammates can reproduce it.

Samples are only valuable if they become a baseline you can fork and evolve.

### FAQ

**Q: Which sample should I start with?**
A: Start with `02-use-cases/beginner/hello_world` to learn the minimal agent + memory loop.

**Q: Why does it use `uv`?**
A: The sample README uses `uv` for dependency management and running (`uv sync`, `uv run agent.py`).

**Q: Is a cloud account required?**
A: Examples may require provider credentials; the Hello World README includes env vars for Volcengine model access.

## Source & Thanks

> Source: https://github.com/bytedance/agentkit-samples
> License: Apache-2.0
> GitHub stars: 330 · forks: 60

---

<!-- ZH -->

## 快速使用

```bash
git clone https://github.com/bytedance/agentkit-samples.git
cd agentkit-samples/02-use-cases/beginner/hello_world

curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv --python 3.12
uv sync

uv run agent.py
```

## 简介

AgentKit Samples 是示例仓库；其中 Hello World 用 `uv` 运行具备短期记忆的对话 Agent，并提供 SSE 流式客户端与环境变量示例，方便本地调试与后续云端部署。

- **适合谁：** 想要可运行的 AgentKit 示例（记忆/工具/MCP），而不是只看概念图的人
- **可搭配：** Python 3.10+；`uv`；示例需要火山引擎凭证接入模型（见 sample README）
- **准备时间：** 15–30 分钟

## 实战建议

- GitHub：330 stars · 60 forks；最近更新 2026-05-12（GitHub API 验证）。
- Hello World 示例 README 用 `uv venv` + `uv sync`，然后 `uv run agent.py` 启动服务。
- 示例包含 SSE 客户端（`client.py`）以及模型/凭证的环境变量配置（见 README）。

## 主要内容

从这个仓库里学到东西的一个高效方法：

1. 先选一个最小用例（比如 `hello_world`）把它端到端跑通。
2. 仔细读用例 README：不少示例包含特定云厂商的凭证与部署步骤。
3. 本地跑通后把它当模板：
   - 先换记忆后端
   - 再加一个工具集成
   - 最后补部署配置
4. 把你用过的命令（`uv venv`、`uv sync`、`uv run …`）记录成团队可复用的 checklist。

示例的价值在于：它能变成团队的可演进基线。

### FAQ

**从哪个示例开始？**
答：建议从 `02-use-cases/beginner/hello_world` 开始，先理解最小的 agent + memory 循环。

**为什么用 `uv`？**
答：示例 README 明确用 `uv` 管依赖与运行（`uv sync`、`uv run agent.py`）。

**一定要云账号吗？**
答：部分示例需要云厂商凭证；Hello World README 给出火山引擎模型访问的环境变量。

## 来源与感谢

> Source: https://github.com/bytedance/agentkit-samples
> License: Apache-2.0
> GitHub stars: 330 · forks: 60


---
Source: https://tokrepo.com/en/workflows/agentkit-samples-hello-world-agent-uv-sse
Author: Agent Toolkit