openviking

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

OpenViking Skill

OpenViking 上下文数据库集成 — 给 AI 朝廷加上长期记忆和知识库。

OpenViking Context Database Integration — Add long-term memory and knowledge base to your AI Court.

什么是 OpenViking

What is OpenViking

OpenViking 是火山引擎开源的 AI Agent 上下文数据库，用文件系统范式统一管理记忆、资源和技能。相比 OpenClaw 默认的 qmd 记忆后端，OpenViking 在大规模文档场景下更强：

能力	qmd（默认）	OpenViking
语义搜索	基础向量匹配	目录递归 + 语义融合
自动摘要	❌	✅ L0/L1/L2 三层
结构化浏览	❌	✅ 虚拟文件系统
Token 节省	❌	✅ 按需加载

OpenViking is an open-source AI Agent context database by Volcano Engine, which unifies the management of memory, resources, and skills using a file system paradigm. Compared to OpenClaw's default qmd memory backend, OpenViking performs better in large-scale document scenarios:

Capability	qmd (Default)	OpenViking
Semantic Search	Basic vector matching	Directory recursion + semantic fusion
Auto-summarization	❌	✅ L0/L1/L2 three layers
Structured browsing	❌	✅ Virtual file system
Token saving	❌	✅ On-demand loading

安装

Installation

1. 安装 Python 包

1. Install Python Package

bash

pip install openviking

bash

pip install openviking

2. 获取 Embedding API Key

2. Get Embedding API Key

推荐使用免费的 NVIDIA NIM API：

访问 https://build.nvidia.com/
登录 → API Keys → 生成 Key
保存 key（以
```
nvapi-
```
开头）

也可以用火山引擎、OpenAI 等其他 provider。

Recommended to use the free NVIDIA NIM API:

Visit https://build.nvidia.com/
Log in → API Keys → Generate Key
Save the key (starts with
```
nvapi-
```
)

You can also use other providers like Volcano Engine, OpenAI, etc.

3. 创建配置文件

3. Create Configuration File

bash

mkdir -p ~/.openviking
cat > ~/.openviking/ov.conf << 'EOF'
{
  "embedding": {
    "dense": {
      "api_base": "https://integrate.api.nvidia.com/v1",
      "api_key": "YOUR_NVIDIA_API_KEY",
      "provider": "openai",
      "dimension": 4096,
      "model": "nvidia/nv-embed-v1"
    }
  },
  "vlm": {
    "api_base": "https://integrate.api.nvidia.com/v1",
    "api_key": "YOUR_NVIDIA_API_KEY",
    "provider": "openai",
    "model": "meta/llama-3.3-70b-instruct"
  }
}
EOF

bash

mkdir -p ~/.openviking
cat > ~/.openviking/ov.conf << 'EOF'
{
  "embedding": {
    "dense": {
      "api_base": "https://integrate.api.nvidia.com/v1",
      "api_key": "YOUR_NVIDIA_API_KEY",
      "provider": "openai",
      "dimension": 4096,
      "model": "nvidia/nv-embed-v1"
    }
  },
  "vlm": {
    "api_base": "https://integrate.api.nvidia.com/v1",
    "api_key": "YOUR_NVIDIA_API_KEY",
    "provider": "openai",
    "model": "meta/llama-3.3-70b-instruct"
  }
}
EOF

4. 设置环境变量

4. Set Environment Variables

bash

echo 'export OPENVIKING_CONFIG_FILE=~/.openviking/ov.conf' >> ~/.bashrc
source ~/.bashrc

bash

echo 'export OPENVIKING_CONFIG_FILE=~/.openviking/ov.conf' >> ~/.bashrc
source ~/.bashrc

使用方式

Usage

Agent 通过 exec 调用

scripts/viking.sh

脚本：

bash

undefined

Agents call the

scripts/viking.sh

script via exec:

bash

undefined

查看状态

Check status

bash skills/openviking/scripts/viking.sh info

索引文件

Index files

bash skills/openviking/scripts/viking.sh add ./my-document.md

批量索引目录

Batch index directory

bash skills/openviking/scripts/viking.sh add-dir ./docs/

语义搜索

Semantic search

bash skills/openviking/scripts/viking.sh search "某个话题"

bash skills/openviking/scripts/viking.sh search "some topic"

浏览已索引的文件

Browse indexed files

bash skills/openviking/scripts/viking.sh list

读取文件摘要

Read file summary

bash skills/openviking/scripts/viking.sh summary <file-path>

undefined

bash skills/openviking/scripts/viking.sh summary <file-path>

undefined

朝廷集成建议

AI Court Integration Suggestions

兵部：索引代码仓库，搜索相关代码片段
户部：索引财务报表，查询历史数据
礼部：索引品牌素材和营销案例
工部：索引运维文档和 runbook
刑部：索引法律法规和合同模板

建议保留 qmd 做日常轻量记忆，OpenViking 做大规模知识库。

Ministry of War: Index code repositories, search related code snippets
Ministry of Revenue: Index financial reports, query historical data
Ministry of Rites: Index brand materials and marketing cases
Ministry of Works: Index operation and maintenance documents and runbooks
Ministry of Justice: Index laws, regulations, and contract templates

It is recommended to keep qmd for daily lightweight memory, and use OpenViking for large-scale knowledge bases.