Skip to main content

OpenEarth

Self-evolving agent system — 让你的 coding agent 越用越强。

CI Release Python

English · 简体中文

介绍

OpenEarth 把 agent 运行过程中的全部经验 —— 轨迹、技能、知识、记忆 —— 沉淀成可检视、可演进、可分享的明文数据,下次会话直接用上,越用越强。

  • 📡 无感采集:后台读各 agent 原生会话库(零插桩),采下完整轨迹(思考 / 工具调用 / 结果),训练与推理共用一套。
  • 🗂️ 经验即数据:轨迹 / 技能 / 知识 / 记忆进同一个明文 git store,可检视、可版本化、可分享。
  • 🌿 从经验演进:在 train/* 分支上从轨迹蒸馏技能,无回退 gate 通过才合并进 main
  • 📚 知识库:文档摄入成可检索 wiki,agent 直接读。
  • 🔌 多 agent:opencode · claude-code · codex · cursor · codeagent。

安装

OpenEarth 有三种角色:

角色 场景 需要 LLM 连接
standalone 个人使用,全部本地闭环 不连接 OpenEarth server
server 接受 client push,提供 pull,承载文档解析 监听 HTTP
client 与 server 同步轨迹 / 知识 / 技能 连 server

把对应模式的 skill 复制给你的 agent,它会自动装好。 详细部署指南见 docs/deployment.md

standalone / server

---
name: openearth-setup
description: 安装并初始化 OpenEarth standalone 或 server 模式。
---

# openearth-setup

**不要索要或回显 API key 明文**,需要就让用户设环境变量。

## 1. 装包

```bash
pip install openearth-ai
```

Windows 如果 Python < 3.10,从 python.org 装新版。Linux (Ubuntu 24.04) 用 venv:
```bash
python3 -m venv ~/.venv-openearth && ~/.venv-openearth/bin/pip install openearth-ai
```

## 2. 设环境变量

**API key 只能放环境变量,不要 echo 明文。**

Windows (PowerShell,重开终端生效):
```powershell
setx OPENEARTH_DOMAIN_AGENT_KEY "sk-..."
```

Linux:
```bash
export OPENEARTH_DOMAIN_AGENT_KEY=sk-...      # 加到 ~/.bashrc 或 ~/.profile
```

## 3. 初始化配置

```bash
openearth config --init
```

编辑 `~/.openearth/config.yaml`,确认 `domain_agent`(`base_url` / `model`),`api_key` 保持 `env:` 形式:

```yaml
domain_agent:
  agent_backend: pydantic-ai
  model: <your-model>
  base_url: https://<your-openai-compatible-endpoint>/v1
  api_key: env:OPENEARTH_DOMAIN_AGENT_KEY
```

> 如果 LLM endpoint 在内网,确认 `no_proxy` 包含其 IP,否则 LLM 请求走 proxy 会被拒绝。

## 4a. standalone:直接启动

```bash
openearth install         # 装技能 + 写 Knowledge pointer + 起 daemon
openearth service status
openearth agents
```

## 4b. server:onboard 后启动

```bash
openearth onboard --role server --host 0.0.0.0 --port <port>
openearth service start
```

`onboard` 会保留上一步的 LLM 配置并打印 **join token**(保存好,client 用它连接)。
daemon 运行 knowledge / trajectory / skills / link 服务。

## 5. 加文档到知识库

```bash
openearth knowledge ingest <path>
openearth knowledge progress
```

client

---
name: openearth-client-setup
description: 安装 openearth client,连接到已运行的 server。
---

# openearth-client-setup

client 不需要 LLM 配置 —— 文档解析在 server 端完成。

## 1. 装包

```bash
pip install openearth-ai
```

## 2. Onboard 为 client

需要 server 的 URL 和 join token:

```bash
openearth onboard --role client \
  --server-url http://<server-ip>:<server-port> \
  --token <join-token> \
  --materialize
```

## 3. 启动 daemon

```bash
openearth service start
openearth service status
```

daemon 自动连接 server、上传轨迹、拉取知识 / 技能、安装到本地 agent(materialize)。

## 4. 提交文档

```bash
openearth doc submit <file> --topic "<topic>"
openearth doc status <job-id>
```

锁版本 / 离线:到 Releases 下载 .whl,pip install openearth_ai-<ver>-py3-none-any.whl

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

openearth_ai-0.4.10.tar.gz (2.4 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

openearth_ai-0.4.10-py3-none-any.whl (1.2 MB view details)

Uploaded Python 3

File details

Details for the file openearth_ai-0.4.10.tar.gz.

File metadata

  • Download URL: openearth_ai-0.4.10.tar.gz
  • Upload date:
  • Size: 2.4 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for openearth_ai-0.4.10.tar.gz
Algorithm Hash digest
SHA256 4e5d20b74edc647258251ce9a22defb74b785d11d10ec2c2cd6ae903eb6870fc
MD5 5a50ce7123ec1f4efebdea0914c94803
BLAKE2b-256 770ceda62e1e757791b97b6f8dffbe4c74ac01ecfc9ce28abe6c5e5cf95b972a

See more details on using hashes here.

File details

Details for the file openearth_ai-0.4.10-py3-none-any.whl.

File metadata

  • Download URL: openearth_ai-0.4.10-py3-none-any.whl
  • Upload date:
  • Size: 1.2 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for openearth_ai-0.4.10-py3-none-any.whl
Algorithm Hash digest
SHA256 158477504e4cdce331397f8a5ba8cf76b1efcc71b8f385cb63acde328a7aea43
MD5 b0095111a9d693a65f0f0c05ebd8a123
BLAKE2b-256 b471bd830653f7d5d3c5bece5b0e0a3d78b397d93489141cfec154fc1f4996ca

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page