Kuku:

Product Introduction

1. Definition: Kuku is a local-first, open-source personal knowledge management (PKM) platform and "second brain" designed specifically for the AI era. Technically, it is a native macOS application built using the Tauri v2 framework and SolidJS, functioning as a high-performance Markdown editor that integrates local indexing with large language model (LLM) capabilities. Unlike traditional note-taking applications that rely on proprietary databases, Kuku operates directly on plain-text Markdown files stored on the user's local file system.

2. Core Value Proposition: Kuku exists to provide a non-proprietary, privacy-centric environment for networked thought. Its primary value lies in transforming static note vaults into dynamic, reusable context for AI. By utilizing "Cursor-style" reviewable diffs and local-first indexing, Kuku allows users to maintain complete data sovereignty while leveraging cutting-edge AI for searching, linking, and editing notes. It targets the "knowledge fragmentation" and "cloud lock-in" problems prevalent in modern productivity stacks.

Main Features

1. Native Local-First Architecture (Tauri & SolidJS): Kuku is engineered for performance, opting for Tauri v2 over the resource-intensive Electron framework. The editor core is built on pure ProseMirror and SolidJS, ensuring a lightweight, "One Light" aesthetic with near-instant load times. Because it is local-first, all Markdown files, wikilinks, and backlinks reside on the user's hardware, ensuring offline functionality and zero-latency file access.

2. Networked Thought Engine (Graph & Bidirectional Linking): The application implements a robust system for connecting ideas, featuring [[wikilinks]] for internal referencing, an automated backlink panel, and a force-directed graph view. This architecture enables users to visualize clusters of information and navigate their knowledge base non-linearly. The indexer ensures that vault paths stay synced with graph nodes, preventing orphaned notes and maintaining structural integrity across thousands of files.

3. AI Agent with Cursor-Style Diff Previews: Kuku integrates an AI agent (currently supporting Google Gemini via "Bring Your Own Key" or BYOK) that can read the entire vault to provide context-aware assistance. Beyond simple chat, the AI can perform complex edits, draft new content based on existing notes, and suggest follow-up prompts. Changes are presented as inline diffs, allowing users to review and approve AI-generated patches before they are committed to the local Markdown source files.

4. Semantic Search and Instant Retrieval: The platform features a local indexing engine that provides full-text search results in under a second. On the roadmap is the transition to hybrid retrieval, combining traditional keyword search with semantic embeddings (powered by local models like nomic-embed-text via Ollama or ONNX Runtime). This allows the AI to "understand" the meaning behind notes rather than just matching characters.

Problems Solved

1. Data Privacy and Vendor Lock-in: Many modern AI note-taking apps require users to upload their entire knowledge base to the cloud, creating privacy risks and making it difficult to switch tools. Kuku solves this by using human-readable Markdown as the source of truth, ensuring the knowledge base remains portable and private.

2. Information Retrieval Friction: As a "second brain" grows, finding specific insights becomes difficult. Kuku addresses this through "Instant Context," where the AI surfaces related notes, links, and past session answers automatically while the user works, preventing the loss of "flow state."

3. Target Audience:

• Software Engineers and Architects: Who require a high-performance, local-first tool for documentation and technical RFCs.
• Researchers and Academic Writers: Who utilize Zettelkasten or networked thought methodologies to link complex ideas.
• Privacy Advocates: Users who demand AI capabilities without sacrificing data sovereignty or being subject to subscription-only cloud models.

4. Use Cases:

• Automated OKR Drafting: Using the AI agent to cross-reference meeting notes and project goals to generate structured planning documents.
• Meeting Intelligence: Utilizing local Whisper.cpp integration to convert voice memos into searchable, linkable Markdown notes.
• Knowledge Synthesis: Using the graph view and semantic search to identify connections between disparate research topics or project backlogs.

Unique Advantages

1. Differentiation: Unlike Obsidian, which is proprietary, or Notion, which is cloud-dependent, Kuku is fully open-source (MIT License) and built for the "AI-native" workflow. While other tools add AI as a sidebar plugin, Kuku treats AI as a core system utility that can proactively edit and organize the vault with human-in-the-loop approvals.

2. Key Innovation: The "Agentic Editing" model is Kuku’s standout innovation. By treating a note vault like a codebase—using diffs and patches—it allows for automated vault maintenance (e.g., fixing broken wikilinks, updating project statuses) that stays grounded in the user's actual writing rather than hallucinating external information.

Frequently Asked Questions (FAQ)

1. Is Kuku a cloud-based or local-only application? Kuku is strictly local-first. Your Markdown files are stored on your machine. While it offers a managed "Pro" sync service for multi-device alignment, the source of truth remains your local files, and the app functions entirely offline.

2. How does Kuku handle AI data privacy? Kuku follows a "Bring Your Own Key" (BYOK) model for Gemini API integration. Requests are routed directly to the provider using your own credentials. For users seeking absolute privacy, the roadmap includes support for local LLMs via Ollama, ensuring no data ever leaves the device.

3. Can I use Kuku with my existing Obsidian or Logseq vault? Yes. Since Kuku uses standard Markdown, wikilinks, and folder structures, you can point Kuku to your existing note folders. It will index the files and provide AI-assisted editing and graph visualization without requiring a proprietary import process.

4. What makes the Kuku editor faster than other note apps? Kuku avoids the overhead of Electron by using Tauri, which leverages the system's native webview. Additionally, the recent rebuild using SolidJS and pure ProseMirror minimizes the JavaScript execution lag often found in complex text editors, resulting in a "lag-free" writing experience even with large files.