Search...Search plugins and themes...
⌘K
Sign in
  • Get started
  • Download
  • Pricing
  • Enterprise
  • Account
  • Obsidian
  • Overview
  • Sync
  • Publish
  • Canvas
  • Mobile
  • Web Clipper
  • CLI
  • Learn
  • Help
  • Developers
  • Changelog
  • About
  • Roadmap
  • Blog
  • Resources
  • System status
  • License overview
  • Terms of service
  • Privacy policy
  • Security
  • Community
  • Plugins
  • Themes
  • Discord
  • Forum / 中文论坛
  • Merch store
  • Brand guidelines
Follow us
DiscordTwitterBlueskyThreadsMastodonYouTubeGitHub
© 2026 Obsidian

Knowledge Topography

joffoojoffoo19 downloads

Interactive semantic topographic contour map for your Obsidian vault. Explore note density, clusters, and peaks as a living landscape.

Add to Obsidian
  • Overview
  • Scorecard
  • Updates4

中文说明 | English Documentation


🇨🇳 中文说明

Knowledge Topography (知识地貌) 是一款专为 Obsidian 打造的语义等高线地形图插件。它通过高维文本嵌入向量(Embeddings)、UMAP 降维流形学习以及二维高斯核密度估计(2D Gaussian KDE),将你的个人笔记库渲染为层叠起伏的 80 层等高线地貌全景图 —— 直观揭示思维的语义聚集、知识山峦(Peaks)与概念峡谷。

✨ 核心特性

  • 🏔️ 80 层微地形等高线:平滑细腻的高阶等高线算法,将抽象的笔记密度转化为生动的地理海拔与连绵山脉。
  • 🎯 语义雷达探针与实时近邻定位 (Radar Probe):当你在 Obsidian 中切换正在编辑的笔记时,地貌地图自动自适应平移聚焦至目标坐标,点亮雷达光斑;底部浮现高质感毛玻璃 HUD 卡片,实时显示所属山峰主题及相似度最高的 Top-3 语义近邻笔记,支持一键点击直接打开。
  • 🧭 自由漫游与全貌回归 (Free Navigation & Reset):全面解锁地图自由平移浏览(Pan),并通过右上角风格胶囊内置的微型回归按钮 ⟲ 或在空白画布双击鼠标瞬间平滑回归完整地貌全景。
  • 🏷️ 自适应山峰识别与 AI 领域命名:采用自适应带宽核密度与局部非极大值抑制(NMS)自动锚定知识高地,并支持联动大语言模型(如 DeepSeek、OpenAI、本地 Ollama 等)自动提炼各山峰的主题领域。
  • 🌐 双语界面原生支持 (i18n):设置后台与视图均提供原生的 简体中文 与 English 双语支持,并支持自动跟随 Obsidian 客户端语言。
  • 🎨 7 款手工定制的美学主题:
    • 宋代青绿山水 (Shanshui):文人界画设色、矿物石青石绿渐变与朱砂印章山标。
    • 普鲁士工程蓝图 (Blueprint):坐标网格底纹、工程图纸准心与严谨测量质感。
    • 深空韦伯星云 (Stellar):深邃宇宙星图感、恒星光芒标记。
    • 瑞士国际主义 (Swiss Minimal):现代主义网格系统、极简黑白方块与理性排版。
    • 和纸苔藓若竹 (Matcha & Bamboo):枯山水苔庭意境、禅意圆环与自然呼吸感。
    • 文艺复兴手稿 (Da Vinci):达芬奇羊皮纸复古棕褐手稿风。
    • 现代极夜深蓝 (Contemporary):合成雷达暗色系与现代数字化界面。
  • ⚡ 灵活的双模向量引擎:
    • 在线接口 (OpenAI 兼容 API):无缝直连本地 Ollama(如 http://localhost:11434/v1)、硅基流动(SiliconFlow)、DeepSeek 或 OpenAI。
    • 本地模型 (完全离线推理):支持本地直接加载 ONNX 嵌入模型(如 bge-m3、jina-embeddings-v2),全程无断网隐忧。
  • 🔌 一键连通性测试:设置面板为向量接口与山峰命名接口均内置了快捷“测试连接”功能,避免因配置错误导致构建失败。
  • 🌐 独立交互式 HTML 导出:可导出为完全自包含的交互式网页,支持脱离 Obsidian 在任意浏览器中独立浏览、平移、缩放与查看笔记卡片。
  • 🔒 严格隐私与零外部遥测:计算全在本地或你指定的 API 服务完成,无任何用户行为跟踪、遥测或第三方云端依赖。

🚀 安装方式

方式一:官方社区插件市场(推荐)

  1. 打开 Obsidian 设置 → 第三方插件。
  2. 关闭 安全模式。
  3. 点击 浏览 并搜索 Knowledge Topography。
  4. 点击 安装 并 启用。

方式二:通过 BRAT 测试版安装

  1. 安装 BRAT 插件。
  2. 在 BRAT 中添加测试仓库:Joffoo/obsidian-knowledge-topography。

方式三:手动安装

  1. 前往最新的 GitHub Releases 页面。
  2. 下载 main.js、manifest.json 与 styles.css 这 3 个文件。
  3. 在你的笔记库中,进入 <配置目录>/plugins/(默认通常为 .obsidian/plugins/)并新建文件夹 knowledge-topography。
  4. 将下载的 3 个文件放入该文件夹中。
  5. 重新加载 Obsidian 并在设置中启用插件。

🛠️ 快速上手与设置说明

  1. 点击左侧边栏的 地图图标,或在命令面板中运行 查看知识地貌 (Open Knowledge Topography)。
  2. 首次使用请进入 插件设置 (Knowledge Topography):
    • 界面语言 (Language):选择 自动跟随系统、简体中文 或 English。
    • 地貌计算 (Topography Engine):
      • 选择 在线接口 (兼容 OpenAI) 并填写服务地址(如 Ollama 或云端 API)与模型名称(如 text-embedding-3-small / bge-m3),轻点 测试连接 确保通畅。
      • 或选择 本地模型 (完全离线) 并指定包含 ONNX 模型的本地文件夹路径。
    • 山峰命名 (Peak Naming):(可选)配置对话大语言模型(如 gpt-4o-mini 或 deepseek-chat)与 API Key,自动提炼山峰领域。
    • 扫描范围 (Scan Scope):按需配置需排除的文件夹(默认忽略 _templates, _attachments 等)与单篇采样字数。
  3. 点击视图中的 构建地貌 按钮。
  4. 地图生成后即可随心探索知识地貌,点击任意散点可直接定位并打开对应笔记。

⚙️ 环境要求

  • Obsidian 版本:v1.8.7 或更高版本。
  • 运行平台:桌面端(macOS、Windows、Linux),因涉及流形降维矩阵运算与本地运行支撑,暂不支持移动端。

🌐 English Documentation

Knowledge Topography transforms your Obsidian vault into an interactive semantic landscape. Using high-dimensional embeddings, UMAP dimensionality reduction, and 2D Gaussian Kernel Density Estimation (KDE), it renders your notes as an 80-level topographic contour map — revealing conceptual mountain ranges, thematic peaks, and hidden valleys of thought.

✨ Features

  • 🏔️ 80-Level Topographic Contours: Generates smooth, high-precision contour lines reflecting note semantic density across your vault.
  • 🎯 Semantic Radar Probe & Real-time Neighbor HUD (Radar Probe): When switching active notes in Obsidian, the topography map smoothly pans and centers on the note's coordinates with an animated radar pulse ripple. A glassmorphic HUD card reveals the summit domain and top-3 closest semantic neighbors with instant click-to-open access.
  • 🧭 Free Navigation & Seamless Reset (Free Pan & Reset): Freely drag and pan across the terrain. Instantly reset to the full global overview by clicking the subtle ⟲ button on the top-right capsule or by double-clicking anywhere on the empty canvas.
  • 🏷️ Intelligent Mountain Peak Detection & AI Naming: Automatically identifies semantic summits using localized non-maximum suppression (NMS) and synthesizes domain labels via LLMs (e.g., DeepSeek, OpenAI, Ollama).
  • 🌐 Native Bilingual Interface (i18n): Fully localized settings and view panels in both English and Simplified Chinese, supporting automatic language detection.
  • 🎨 7 Handcrafted Aesthetic Themes:
    • 宋代青绿山水 (Shanshui): Traditional Chinese landscape painting with mineral pigments and red seal badges.
    • 普鲁士工程蓝图 (Blueprint): Technical precision cyanotype with technical crosshair markers.
    • 深空韦伯星云 (Stellar): Deep cosmic nebula survey aesthetic with stellar marks.
    • 瑞士国际主义 (Swiss Minimal): Helvetica-inspired modernist layout with clean typographic boxes.
    • 和纸苔藓若竹 (Matcha & Bamboo): Zen garden atmosphere with subtle matcha green gradients.
    • 文艺复兴手稿 (Da Vinci): Renaissance parchment with vintage sepia tonality.
    • 现代极夜深蓝 (Contemporary): Sleek synthetic radar dark mode.
  • ⚡ Flexible Embedding Engines:
    • Online API (OpenAI Compatible): Seamlessly connect to Ollama (local http://localhost:11434/v1), SiliconFlow, DeepSeek, or OpenAI endpoints.
    • Local Model (Fully Offline): Run completely offline using local ONNX embeddings (e.g., bge-m3, jina-embeddings-v2).
  • 🔌 Built-in Connection Diagnostics: Test your API configuration with a single click before initiating full builds.
  • 🌐 Interactive Standalone HTML Export: Export a self-contained, publication-ready interactive HTML file that can be opened in any modern browser without Obsidian running.
  • 🔒 Zero Telemetry & Strict Privacy: All processing runs locally within your vault or connects solely to your designated API endpoint. No external user tracking or telemetry.

🚀 Installation

Via Community Plugins (Recommended)

  1. Open Obsidian Settings → Community plugins.
  2. Turn off Restricted mode.
  3. Click Browse and search for Knowledge Topography.
  4. Click Install, then Enable.

Via BRAT

  1. Install the BRAT plugin.
  2. Add beta plugin: Joffoo/obsidian-knowledge-topography.

Manual Installation

  1. Download main.js, manifest.json, and styles.css from the latest GitHub Release.
  2. In your vault, navigate to <config-dir>/plugins/ (usually .obsidian/plugins/) and create a folder named knowledge-topography.
  3. Place the downloaded files inside that folder.
  4. Reload Obsidian and enable the plugin in settings.

🛠️ Getting Started & Configuration

  1. Click the Map icon on the left ribbon or run the command Open Knowledge Topography (查看知识地貌).
  2. Go to Plugin Settings (Knowledge Topography):
    • Language: Choose Auto (System Default), English, or 简体中文.
    • Topography Engine:
      • Choose Online API (OpenAI Compatible) and enter your server URL (e.g. Ollama or cloud endpoint) and embedding model name (e.g., text-embedding-3-small), then click Test Link.
      • Or choose Local Model (Offline) and specify the local directory containing your ONNX model.
    • Peak Naming: (Optional) Configure a chat completion model (e.g. gpt-4o-mini or deepseek-chat) to automatically generate summits' domain names.
    • Scan Scope: Customize folders to skip (defaults to _templates, _attachments) and per-note sample sizes.
  3. Click Build Topography in the view panel.
  4. Enjoy exploring your knowledge landscape! Click on any note point to jump directly to the corresponding note.

⚙️ Requirements & Compatibility

  • Obsidian: v1.8.7 or higher.
  • Platform: Desktop only (macOS, Windows, Linux). Requires desktop environment for spatial matrix computations and local rendering.

📄 License

This project is licensed under the MIT License.

HealthExcellent
ReviewSatisfactory
About
Visualize your vault as an 80-layer semantic contour map using high-dimensional embeddings, UMAP and 2D Gaussian KDE to reveal semantic clusters, knowledge peaks and concept valleys. Export interactive standalone HTML, auto-identify peaks with optional LLM summaries, run embeddings locally (ONNX) or via OpenAI/Ollama, and keep all processing private with no external telemetry.
VisualizationAIExport
Details
Current version
0.1.5
Last updated
4 hours ago
Created
Yesterday
Updates
4 releases
Downloads
19
Compatible with
Obsidian 1.8.7+
Platforms
Desktop only
License
MIT
Report bugRequest featureReport plugin
Author
joffoojoffoo
GitHubjoffoo
  1. Community
  2. Plugins
  3. Visualization
  4. Knowledge Topography

Related plugins

Mermaid View

Handle Mermaid files as first-class citizens, rendering them fullscreen when opened.

Markmind

Mind map, outline and PDF annotation tool. (Closed source)

Vault Graph

Your whole vault as one disc: notes packed into folder wedges on a fixed lattice, sized by links, with a growth heatmap, timeline, search and per-folder filters. Deterministic, not force-directed - the same vault always draws the same picture.

NanoBanana PRO

Generate Knowledge Posters (infographics) from your notes using AI. Supports OpenAI, Gemini, Anthropic, and xAI.

Excalidraw

Visual PKM powerhouse. Create and edit Excalidraw drawings.

Advanced Canvas

Supercharge your canvas experience. Create presentations, flowcharts and more.

Copilot

Run AI agents such as Claude Code, Codex, and OpenCode inside your vault. Turn your second brain into a smart assistant that gets knowledge work done.

ExcaliBrain

An interactive, structured mind-map of your Obsidian vault.

Claudian

Embeds Claude Code/Codex and other local Agents as AI collaborators in your vault.

Smart Connections

Find related notes and excerpts while writing. Your AI link building copilot displays relevant content in graph + list view. A local embedding model powers semantic search. Zero setup. No API key.