AI Skill Hub 强烈推荐:kektordb MCP工具 是一款优质的AI工具。AI 综合评分 8.2 分,在同类工具中表现稳健。如果你正在寻找可靠的AI工具解决方案,这是一个值得深入了解的选择。
kektordb MCP工具 是一款基于 Go 开发的开源工具,专注于 AI记忆系统、向量数据库、知识图谱 等核心功能。作为 GitHub 开源项目,它拥有活跃的社区支持和持续的版本迭代,代码完全透明可审计,支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流,都能提供稳定可靠的解决方案。
kektordb MCP工具 是一款基于 Go 开发的开源工具,专注于 AI记忆系统、向量数据库、知识图谱 等核心功能。作为 GitHub 开源项目,它拥有活跃的社区支持和持续的版本迭代,代码完全透明可审计,支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流,都能提供稳定可靠的解决方案。
# 方式一:go install(推荐) go install github.com/sanonone/kektordb@latest # 方式二:从源码编译 git clone https://github.com/sanonone/kektordb cd kektordb go build -o kektordb . # 方式三:下载预编译二进制 # 访问 Releases 页面下载对应平台二进制文件 # https://github.com/sanonone/kektordb/releases
# 查看帮助 kektordb --help # 基本运行 kektordb [options] <input> # 详细使用说明请查阅文档 # https://github.com/sanonone/kektordb
# kektordb 配置说明 # 查看配置选项 kektordb --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export KEKTORDB_CONFIG="/path/to/config.yml"
The cognitive memory layer for AI agents.
<p align="center"> <img src="docs/images/logo.png" alt="KektorDB Logo" width="500"> </p>
<p align="center"> <a href="DOCUMENTATION.md">Documentation</a> • <a href="CONTRIBUTING.md">Contributing</a> • <a href="docs/guides/zero_code_rag.md">GraphRAG Guide</a> </p>
KektorDB is an AI memory system - not a database that stores data, but an engine that understands what it stores. It combines high-performance vector search (HNSW) with a temporal knowledge graph and a cognitive engine that continuously analyzes your data, detects contradictions, tracks importance, and lets irrelevant information fade naturally.
<p align="center"> <img src="docs/images/kektordb-demo.gif" alt="KektorDB Demo" width="800"> </p>
In 30 seconds - give your AI agent persistent memory:
```bash
profile = client.get_user_profile("user_42", "agent_memory") print(f"Style: {profile.get('communication_style')}") ```
kektordb setup opencode
./kektordb try
Starts an ephemeral demo server on http://127.0.0.1:9091 with seeded sample memories, one entity, and graph links. All data lives in a temp directory and is discarded on exit. Vector search works out of the box; text search (query_text) also works when an embedder is available (Ollama or the built-in ONNX runtime):
curl -X POST http://127.0.0.1:9091/vector/actions/search-with-scores \
-d '{"index_name":"mcp_memory","k":5,"query_text":"how does the gardener work?"}'
docker pull none99/kektordb docker run -p 9091:9091 -v $(pwd)/data:/data none99/kektordb ```
---
| Mode | Command | Best for |
|---|---|---|
| **MCP Server** | kektordb --mcp --tools=agent | AI agent memory (Claude, Cursor, Codex, Gemini CLI, OpenCode) |
| **REST Server** | ./kektordb | HTTP API backend, any language |
| **Go Library** | import "github.com/sanonone/kektordb/pkg/engine" | Embedded in-process, zero network overhead |
| **AI Gateway** | ./kektordb -enable-proxy -proxy-config=proxy.yaml | Zero-code RAG between Chat UI and LLM |
| **Python/TS Client** | pip install kektordb-client | Application integration |
---
KektorDB includes an optional built-in ONNX embedder (all-MiniLM-L6-v2, 384 dimensions) powered by Rust/Candle for zero-config local embeddings - no Ollama required.
Build with Rust support:
make build-rust-native # requires protoc (auto-downloaded by Makefile)
make run-rust
The ONNX model (~90 MB) is downloaded automatically from HuggingFace on first launch with SHA256 verification.
| Mode | Description |
|---|---|
auto | Auto-detect: local ONNX if available, Ollama as fallback (default) |
ollama / ollama_api | Use Ollama embedding API |
openai / openai_compatible | Use OpenAI-compatible embedding API |
gemini / google | Use Gemini embedContent API |
local | Built-in ONNX model (requires -tags rust build) |
---
```python from kektordb_client import KektorDBClient from kektordb_client.cognitive import CognitiveSession
client = KektorDBClient(port=9091) client.vcreate("agent_memory", metric="cosine")
创新的向量+时间图谱混合架构,填补AI记忆系统空白。文档完善,社区活跃,Go实现性能稳定,值得MCP生态关注。
该工具使用 NOASSERTION 协议,商用场景请仔细阅读协议条款,必要时咨询法律意见。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
📄 NOASSERTION — 请查阅原始协议条款了解具体使用限制。
总体来看,kektordb MCP工具 是一款质量优秀的AI工具,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | kektordb |
| 原始描述 | 开源MCP工具:AI memory system combining vector search with temporal knowledge graph. Built-in。⭐73 · Go |
| Topics | AI记忆系统向量数据库知识图谱MCP工具Go语言 |
| GitHub | https://github.com/sanonone/kektordb |
| License | NOASSERTION |
| 语言 | Go |
收录时间:2026-05-17 · 更新时间:2026-05-19 · License:NOASSERTION · AI Skill Hub 不对第三方内容的准确性作法律背书。