The LLM Knowledge Base your AI can read and write.
Not another note app. A protocol-native knowledge base where Claude, ChatGPT, Cursor, and any MCP client can save, search, and grow your long-term knowledge — in the middle of a conversation.
Voice works too — record a note, AI transcribes, extracts insights, and saves to your knowledge base.
Your memory. Your knowledge. Truly yours — portable across every AI you use.
Works natively with
The Problem
Your Knowledge Is Leaking Away
AI made you 10x faster at generating knowledge — but your way of saving and finding it is still stuck 10 years ago.
Your Knowledge Is Locked Inside One LLM
The decisions, preferences, and insights you've built up in Claude? Switch platforms and they're gone. Account banned, platform shuts down, or you move to GPT-5 — your hard-earned AI context vanishes. Start from scratch.
Manual Organization Is Exhausting
Open Notion → pick a folder → write a title → add tags → paste content. 5 minutes per knowledge entry, 10 entries a day = 50 minutes — pure manual labor.
Saved but Can't Find It
"I noted something about user growth earlier..." Search 'growth', get 30 results, none of them right. Because you wrote it as 'PMF validation strategy'.
How it works
Three Steps to a Living LLM Knowledge Base
From input to retrieval, fully AI-driven. You just focus on thinking.
Frictionless Input
Just say it naturally
Say 'remember this' in any AI chat, or let MCP auto-capture. Zero friction, zero extra steps.
AI Auto-Structures It
No manual categorization
AI automatically extracts titles, categories, tags, and vectorizes. Every piece of knowledge is indexed and ready to retrieve.
Smart Retrieval
Semantic search, not keywords
Vector semantic search — AI finds what's truly relevant. No need to remember keywords, just describe what you're looking for.
Features
A Knowledge Base Redesigned for the AI Era
Not AI bolted onto an old tool — built from scratch for AI collaboration. Your knowledge and memory, always yours.
MCP Native Integration
Read and write knowledge directly in AI chats
Deep integration with Claude Desktop / Web / Code via MCP protocol. No app switching — just say 'remember this' in any AI conversation. Configure once, share across all AI clients.
Voice to Knowledge
One recording replaces manual notes
Upload audio or share from your phone. AI handles three-layer processing: ASR transcription → spoken language cleanup → structured extraction. Commute ideas, meeting notes, book summaries — speak and it's saved.
AI Auto-Structuring
Zero manual categorization
When saving knowledge, AI auto-extracts titles, classifies type (note/insight/tech log/article/reflection), generates tags, and vectorizes. New knowledge auto-discovers and links related existing knowledge.
Daily Review & AI Insights
Knowledge isn't just about saving
GitHub-style contribution heatmap tracks your knowledge input. AI recommends what to review based on vector similarity, generating full knowledge base insight reports — find blind spots, identify trends.
LLM-Native Memory Layer
Your memory, always with you
Through save_memory / recall_memory / get_soul MCP tools, Claude auto-extracts decisions, insights, and lessons during conversations and saves them to your knowledge base. 0.90 vector dedup, auto-categorization, Soul profile export. Switch AI platforms? Memory stays.
Bilingual Interface
Switch between English and Chinese instantly
Full support for English and Chinese UI with URL-based locale routing (/zh/ and /en/). Auto-detects your browser language on first visit. Manual toggle in the top-right corner. Your knowledge base content stays intact across language switches — AI search works regardless of interface language.
AI-Native Architecture
Not a Note App with AI Bolted On
A Knowledge Base Built for AI
AI-native from the ground up: vector indexing, semantic understanding, auto-structuring — every layer designed for AI collaboration.
Vector Semantic Search
Describe your question in natural language, AI pinpoints the answer. No need to remember keywords — 'that counterintuitive insight about growth' still works.
Semantic understanding, not keyword matchingAI Auto-Tagging & Classification
When saving knowledge, AI automatically extracts titles, classifies type (note / insight / tech log / article / reflection), and generates precise tags. You just write, AI handles the rest.
5 knowledge types auto-detectedRelated Knowledge Auto-Discovery
When new knowledge is saved, it automatically discovers and links to similar existing knowledge. No manual linking needed — your knowledge network grows on its own.
Auto-linked when similarity threshold metAI Insight Reports
Analyzes your full knowledge base to identify blind spots, high-frequency themes, and evolution trends. Not just storage — helps you understand your own knowledge structure.
Daily review + smart recommendations + heatmapMemory Dedup & Reinforcement
Same content won't be saved twice — when vector similarity ≥ 0.90, existing memory is reinforced with updated count and timestamp. The more important a decision, the more reinforced and weighted in search.
0.90 cosine similarity thresholdSoul Profile
Auto-distills your user profile from accumulated memories: background, thinking patterns, work preferences. Export as a System Prompt and paste into any AI platform — it understands you from the first message.
Exportable as cross-platform System PromptUse Cases
Who Uses KnowMine?
Anyone who works with AI needs a knowledge base that keeps up with AI's pace.
Indie Developers
Debug experience never lost again
Spent 3 hours fixing a weird Docker networking issue. Told Claude Code 'remember this debugging process'. Three months later, hit a similar problem — Claude pulled up the solution immediately.
Heavy AI Users
Conversation insights preserved forever
Every day chatting with Claude about architecture, business analysis, creative ideas — generating tons of valuable insights. Before, closing the window meant losing it all. Now via MCP, it's auto-saved. Next conversation, AI connects your past thinking.
Content Creators
Inspiration ready when you need it
Spotted an interesting idea, jotted down 'counterintuitive insight about XX trend'. Two weeks later writing an article, searched 'trend' — semantic match found it + 5 related pieces. Article framework assembled itself.
Lifelong Learners
Knowledge truly internalized
Finished a great book, entered core ideas into KnowMine. Daily review uses vector similarity to suggest what to revisit, AI insights uncover hidden connections between different books.
Testimonials
What Users Say
“Finally no more switching between Notion and Claude. Say 'remember this' in the conversation, and it's saved automatically.”
“The voice-to-knowledge feature is incredibly practical. Record during my commute, arrive at the office and it's already a structured note.”
“After three months, the most surprising thing is the auto-linking — it discovered connections between my knowledge that I hadn't even noticed myself.”
Blog
Latest Articles
KnowMine Claude Code Plugin: Your Knowledge Base, Inside Every Session
KnowMine now has an official Claude Code Plugin. One command to install — Claude Code automatically gets your personal knowledge base. Save during conversations, recall across sessions, never re-explain context again.
One Knowledge Base, Any AI: How to Stop Rebuilding Context Every Time You Switch Models
Every time you open a new AI chat, it forgets everything. The real cost isn't the switch — it's the 10-15 minutes rebuilding context. Here's how an MCP-native knowledge base solves model lock-in for good.
The Word Karpathy Used Was 'Compile' — And That Changes Everything
Everyone is copying Karpathy's three folders. Almost no one is paying attention to the verb he used to describe what happens between them. The verb is the whole thesis.
MCP Native
Knowledge That Moves with AI
Deep integration with all major AI clients via MCP (Model Context Protocol). Not just an API connection — native protocol-level fusion. Your knowledge is at hand in every AI conversation.
Claude Desktop
LiveNative desktop integration
ChatGPT
LiveOAuth auto-authorization
Claude Code
LiveDeep CLI integration
Qwen Chat
LiveDesktop app with MCP support
Cursor / VS Code
LiveIDE integration
More Platforms
LiveOpenClaw and other MCP clients
Why MCP?
MCP (Model Context Protocol) is Anthropic's open protocol standard that lets AI models securely access external data sources. As an MCP Server, KnowMine natively supports all MCP clients.
{
"mcpServers": {
"knowmine": {
"url": "https://knowmine.ai/api/mcp"
}
}
}Pricing
Simple, Transparent Pricing
Free plan covers daily use, Pro unlocks everything
Free
Start building your LLM knowledge base
- 100 knowledge entries
- AI auto-classification & tagging
- Vector semantic search
- MCP integration (1 client)
- Daily review
- Knowledge graph visualization
Pro
Designed for heavy knowledge workers
- Unlimited knowledge storage
- AI auto-classification & tagging
- Advanced vector semantic search
- MCP integration (unlimited clients)
- Daily review + AI insight reports
- Knowledge graph visualization
- Knowledge spaces (domain management)
- Priority technical support
Start Building Your LLM Knowledge Base
Get started free, no credit card required. Connect your AI clients and let knowledge truly work for you.
Get Started Free