Personal RAG and Semantic Search from AI Chat
Create the AI
version of you.
Your knowledge, stored and available in every conversation.
Your knowledge, stored and available in every conversation.
Voice note, document, URL, thought. Captured in seconds, tagged automatically. You are doing the work anyway. Now you are capturing it too.
Ask what your principles are. What your voice sounds like. What you said last month. Get it back in your own words, source always cited.
Everything AI creates starts from your knowledge, your taste, your judgment. Not generic. Not almost right. Exactly you.
Every idea, principle, decision, and document you care about. Stored the moment it arrives. From a voice note in a taxi to a URL you found at midnight.
add_knowledge
add_voice_note
add_from_url
add_document
How you like to run meetings. What you decided last month. The way you explained that concept to a client. Search semantically and it comes back in seconds, in your words, source cited.
search_twin
get_by_tag
get_by_type
list_recent
get_sources
Ask your twin to synthesise everything you know on a topic. Ask it to find the patterns you repeat without naming. Over time it surfaces the principles you hold but never wrote down.
synthesise
find_patterns
get_schema
Every note you store, every search you run, every pattern you surface. You are building something.
Not just a knowledge base. A digital twin. Your judgment, your voice, your taste, your principles. Available in every conversation, forever. Most people's best thinking disappears. Yours compounds.
RAG stands for Retrieval Augmented Generation. It is the architecture that lets AI answer questions using your specific knowledge rather than its training data alone. Until now, building one required an engineer, a vector database, an embedding pipeline, and a weekend. MyAITwin gives you all of that from a chat message. Every note you store is embedded and indexed in Pinecone, a production-grade vector database.
Ask "how do I approach difficult conversations?" and it finds your relevant notes even if you never used those exact words. It understands intent, not just text.
Keyword search finds what you typed. Semantic search finds what you meant.
No dashboards. No upload flows. No configuration. Your RAG database grows every time you share something worth keeping in conversation.
Powered by Pinecone. Production infrastructure. Zero setup.
Define your own knowledge types. Add your own tags. Edit anything as your thinking evolves. Delete anything that no longer holds. The structure reflects how you think, not how someone else decided you should organise.
add_schema_type
update_schema_type
update_knowledge
delete_knowledge
You will need MCP capabilities to get started.
Requires Claude (Pro, Team, or Max) with connectors enabled.