Summary
A personal RAG system that turns your chats into a searchable knowledge base where everything sounds like you when retrieved. You get 19 tools split between storing knowledge (add_knowledge, add_voice_note, add_from_url), retrieving it with semantic search (search_twin, get_by_type), and managing schema (add_schema_type). Runs on PostgreSQL for structured data and Pinecone for vector search. The architectural choice that matters: it separates what you know from how you say things, so when you ask for output it combines both. OAuth 2.1 authentication, multi-tenant isolation tested with 35 cross-tenant checks, and you control the type system from inside Claude. Free during early access at myaitwin.lutolearn.com.
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
Tools
Public tool metadata for what this MCP can expose to an agent.
19 tools
add_knowledgeStore a piece of knowledge or a skill in your twin. Automatically tagged and embedded for search. The twin stores two fundamentally different things — knowledge (what the user knows: transcripts, decisions, ideas, observations) and skills (how they express things: their Linked...6 params
Store a piece of knowledge or a skill in your twin. Automatically tagged and embedded for search. The twin stores two fundamentally different things — knowledge (what the user knows: transcripts, decisions, ideas, observations) and skills (how they express things: their Linked...
Parameters* required
tagsarrayManual tags to add on top of auto-generated ones
typestringType for this item. Common types:
• skill — how you express something. Your LinkedIn voice, your email style, your proposal structure. The craft layer that shapes knowledge into output.
• principle — repeating values, rules, guidelines you apply consistently.
• knowledge — expertise areas, domain knowledge, what you know deeply.
• idea — concepts, hypotheses, things you're exploring.
• voice — writing style, tone, how you communicate.
• brand — visual preferences, aesthetic principles, brand rules.
• template — reusable structures, formats, scaffolding.
• resource — links, documents, references you trust.
• reference-record — created via add_reference_record after a creation task. Do not store directly.
• meta-principle — surfaced by find_patterns after enough reference records exist.
Or any custom type already in the user's schema.
titlestringShort label for this item (optional)
contentstringThe actual content — write it clearly, in the user's voice
provenancestringWhere this content originates. personal = the user's own thinking. organisational = from their organisation (e.g. team docs, internal). external = from outside (articles, reports, third-party authors). Default: personal.one of
personal · organisational · externalsource_refstringWhere this came from (document name, URL, etc.)
add_reference_recordRecord a creation event after the user has produced something with the twin's help — pairs knowledge used + skill applied + output produced + the nuance of this specific case. The system prompt instructs you to offer this after significant creation tasks; call it once the user...6 params
Record a creation event after the user has produced something with the twin's help — pairs knowledge used + skill applied + output produced + the nuance of this specific case. The system prompt instructs you to offer this after significant creation tasks; call it once the user...
Parameters* required
tagsarrayTags — inherit from the knowledge and skill plus any case-specific tags
titlestringShort description of what was created (e.g. "LinkedIn post about Q3 outlook")
nuancestringWhat was different or adapted in THIS specific case — the lesson worth keeping
skill_idstringID of the skill applied (from search_for_creation skills bucket)
knowledge_idsarrayIDs of knowledge items used (from search_for_creation results)
output_summarystringOne-paragraph summary of the output that was produced
add_from_urlFetch a URL, extract what's worth knowing, and store it. The twin analyses the page against your existing schema and reports what it found before confirming what was stored.2 params
Fetch a URL, extract what's worth knowing, and store it. The twin analyses the page against your existing schema and reports what it found before confirming what was stored.
Parameters* required
urlstringThe URL to fetch and analyse
notesstringWhy you're ingesting this — adds context to what gets stored
add_documentIngest a document by providing its text content. Chunks and stores with source tracking. Good for PDFs, notes, reports.3 params
Ingest a document by providing its text content. Chunks and stores with source tracking. Good for PDFs, notes, reports.
Parameters* required
notesstringContext about this document
contentstringThe full text content of the document
filenamestringDocument name (used as source reference in all future results)
add_voice_noteIngest a voice note transcript. The twin extracts principles, decisions, and ideas — identifying what you said that's worth keeping, in your language. Source is tracked as voice note + date.3 params
Ingest a voice note transcript. The twin extracts principles, decisions, and ideas — identifying what you said that's worth keeping, in your language. Source is tracked as voice note + date.
Parameters* required
datestringDate of the voice note, e.g. "2026-05-16"
notesstringWhat this voice note was about
transcriptstringThe full transcript text from Otter or similar
get_schemaShow what knowledge types exist in your twin, how many items of each type are stored, and the total knowledge base size.
Show what knowledge types exist in your twin, how many items of each type are stored, and the total knowledge base size.
No parameter schema in public metadata yet.
add_schema_typeAdd a new knowledge type. Do this from Claude chat — no database access needed. The type is immediately available for storing knowledge.2 params
Add a new knowledge type. Do this from Claude chat — no database access needed. The type is immediately available for storing knowledge.
Parameters* required
namestringType name — lowercase, short, e.g. "case-study" or "objection"
descriptionstringWhat this type stores — one clear sentence
update_schema_typeUpdate the description of an existing knowledge type.2 params
Update the description of an existing knowledge type.
Parameters* required
namestringThe type name to update
descriptionstringNew description
search_twinSemantically search everything in your twin. Returns the most relevant items ranked by relevance, each with its source cited.3 params
Semantically search everything in your twin. Returns the most relevant items ranked by relevance, each with its source cited.
Parameters* required
typestringFilter to a specific knowledge type
querystringWhat you're looking for — natural language
top_knumberHow many results to return (default 10, capped at 10)
get_by_typeRetrieve all knowledge of a specific type, newest first.2 params
Retrieve all knowledge of a specific type, newest first.
Parameters* required
typestringKnowledge type to retrieve
limitnumberMax items to return (default 20)
get_by_tagRetrieve all knowledge with a specific tag.2 params
Retrieve all knowledge with a specific tag.
Parameters* required
tagstringTag to filter by
limitnumberMax items to return (default 20)
search_for_creationUse this BEFORE any creation task ("help me write X", "I'm working on Y"). Runs two parallel searches and returns them separately: a SKILLS bucket (skill/voice/template — the craft layer) and a KNOWLEDGE bucket (knowledge/principle/brand/idea/resource — the material). Bring bo...2 params
Use this BEFORE any creation task ("help me write X", "I'm working on Y"). Runs two parallel searches and returns them separately: a SKILLS bucket (skill/voice/template — the craft layer) and a KNOWLEDGE bucket (knowledge/principle/brand/idea/resource — the material). Bring bo...
Parameters* required
querystringWhat you're looking for — natural language, the task or topic at hand
output_typestringWhat kind of output the user is creating (e.g. "linkedin-post", "client-proposal", "follow-up-email"). Used to track the skill gap if no matching skill exists.
synthesisePull everything relevant on a topic and return structured, usable output — ready to turn into a brief, a slide, or a team handoff. Every source cited.3 params
Pull everything relevant on a topic and return structured, usable output — ready to turn into a brief, a slide, or a team handoff. Every source cited.
Parameters* required
typestringFocus synthesis on a specific knowledge type
top_knumberHow many knowledge items to draw from (default 10, capped at 10)
topicstringThe topic or question to synthesise on
find_patternsAnalyse your stored knowledge and surface recurring themes — principles you repeat without naming, patterns in how you think. Especially powerful after adding several voice notes.1 params
Analyse your stored knowledge and surface recurring themes — principles you repeat without naming, patterns in how you think. Especially powerful after adding several voice notes.
Parameters* required
focusstringFocus analysis on a specific type, e.g. "principle" or "voice-note"
get_sourcesShow all documents, URLs, and voice notes that have been ingested — with what was extracted from each.1 params
Show all documents, URLs, and voice notes that have been ingested — with what was extracted from each.
Parameters* required
source_typestringFilter by source typeone of
document · url · voice-notelist_recentShow the most recently added knowledge items.2 params
Show the most recently added knowledge items.
Parameters* required
typestringFilter to a specific type
limitnumberHow many to show (default 10)
update_knowledgeEdit a stored knowledge item. Provide the item ID and the fields to change.6 params
Edit a stored knowledge item. Provide the item ID and the fields to change.
Parameters* required
idstringThe knowledge item ID (from search_twin or list_recent)
tagsarrayReplace tags
typestringChange the type
titlestringNew title
contentstringNew content (will be re-embedded)
provenancestringUpdate provenance: personal / organisational / externalone of
personal · organisational · externaldelete_knowledgeRemove a knowledge item from your twin. This is irreversible.1 params
Remove a knowledge item from your twin. This is irreversible.
Parameters* required
idstringThe knowledge item ID to delete
get_welcomeReturns the user's personalised session opener — a short message that surfaces twin state, recent activity, and current focus. Call this on the user's first message each session, before responding to anything else.
Returns the user's personalised session opener — a short message that surfaces twin state, recent activity, and current focus. Call this on the user's first message each session, before responding to anything else.
No parameter schema in public metadata yet.
MyAITwin MCP
A personal RAG database and semantic search engine you build and control from inside your AI chat. Store your knowledge, voice, and skills as you work. Retrieve them in seconds, source always cited. Your AI then creates output that is recognisably you, in every conversation.
Live at https://myaitwin.lutolearn.com. Free during early access.
What it is
MyAITwin is two things at once.
The toolbox. A production-grade RAG database with semantic search that you shape from chat. You define the structure, the types, the tags. It is yours, it is visible, and you are the architect of it.
The twin. The layer on top that greets you, guides you, assesses what you store, and creates output that sounds like you. It knows the difference between what you know (your knowledge) and how you say things (your skills), and it uses both.
Install
Three steps. Under two minutes.
- Sign up at https://myaitwin.lutolearn.com/ with your email.
- Click the magic link, then copy your personal MCP URL from
/create. - In Claude Desktop: Settings → Connectors → Add custom connector → paste your URL.
Or use the canonical OAuth-authenticated endpoint:
- URL:
https://myaitwin.lutolearn.com/mcp - Transport: Streamable HTTP
- Auth: OAuth 2.1 with PKCE (S256) and Dynamic Client Registration
Requires a client with MCP capability. Currently Claude Pro, Claude Team, and ChatGPT Pro.
The 19 tools
Storing knowledge
| Tool | What it does |
|---|---|
add_knowledge | Store a typed, tagged knowledge item |
add_voice_note | Store a voice note transcript with automatic extraction |
add_document | Store a long document with automatic chunking |
add_from_url | Fetch and store a web page |
add_reference_record | Store a creation event linking knowledge and skills used |
Retrieving knowledge
| Tool | What it does |
|---|---|
search_twin | Semantic search across all knowledge |
search_for_creation | Dual search returning skills and knowledge separately |
get_by_type | Retrieve all items of a specific type |
get_by_tag | Retrieve all items with a specific tag |
list_recent | List recently added items |
Understanding your twin
| Tool | What it does |
|---|---|
get_schema | Overview of your types and how many items you have |
get_sources | List all source documents |
find_patterns | Surface recurring patterns across your knowledge |
synthesise | Synthesise across multiple knowledge items on a topic |
Managing your twin
| Tool | What it does |
|---|---|
get_welcome | Session initialisation and system prompt |
update_knowledge | Update an existing item |
add_schema_type | Define a new knowledge type |
update_schema_type | Update an existing type definition |
delete_knowledge | Delete an item (destructive) |
All tools are annotated with title, readOnlyHint, and destructiveHint
per the MCP spec. Of the 19: 10 read-only, 8 write (non-destructive), 1
destructive (delete_knowledge).
How it works
RAG is Retrieval-Augmented Generation. It is the architecture that lets AI answer using your specific knowledge rather than its training data alone.
Two layers:
- Supabase (PostgreSQL) for structured records with types, tags, and provenance.
- Pinecone for vector embeddings, so you can search by meaning rather than exact words.
When you search, both layers work together and return results ranked by relevance. Every result is cited with source and date, and tagged with provenance: personal (your own thinking), organisational (from your organisation), or external (from someone else).
The architectural insight worth getting right:
Knowledge is what you know. Facts, decisions, transcripts, observations.
Skills are how you express things. Your LinkedIn voice. Your email style. Your proposal structure.
Exceptional output needs both. Take a meeting transcript and ask for a follow-up email. The twin needs the transcript and your email skill to produce something that is accurate and unmistakably yours. Neither alone is enough.
Security and privacy
- Bearer token authentication on every request, hashed at rest.
- OAuth 2.1 with PKCE for connector-style integration. No shared secrets, no static credentials.
- Multi-tenant data isolation: each user lives in their own namespace. Other users can never read your data. Verified by a 35-check cross-tenant test suite.
- Rate limiting per tenant.
- Append-only audit log on every tool call.
- Prompt injection guardrails on stored content.
- Your data is used only to provide the service. Never used to train AI models. Never shared with third parties.
- You can delete your account and all data instantly from
/create. Deletion is immediate and irreversible.
Privacy policy: https://myaitwin.lutolearn.com/privacy Security contact: security@lutolearning.com Privacy contact: privacy@lutolearning.com
Distribution
- Official MCP Registry:
com.lutolearn/myaitwin - Anthropic Connectors Directory: submitted, in review
- Listed at: Glama, mcp.so, mcp.directory, mcpserverfinder, Hugging Face, awesome-mcp-servers
License
MIT. See LICENSE.
Links
- Website: https://myaitwin.lutolearn.com
- Documentation: https://myaitwin.lutolearn.com/docs
- Privacy: https://myaitwin.lutolearn.com/privacy
- Support: support@lutolearning.com
MyAITwin MCP by Luto.
Featured
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →Registryactive
TransportHTTP
UpdatedMay 24, 2026
