AI Twin

A private, typed memory layer that sits above your AI assistant.

Beginner
Screenshot of AI Twin, A private, typed memory layer that sits above your AI assistant

What is AI Twin?

AI Twin is a UK-hosted, model-agnostic memory layer for personal and professional life. You upload documents, notes, voice memos and exports, confirm what is kept, and query a typed, audited memory that plugs into assistants such as Claude via MCP connectors.

AI Twin is a personal AI memory layer built by a UK company. Rather than keeping what it learns about you inside a single chatbot, it stores your context in its own governed database and lets you pull that memory into whichever assistant you choose. You decide what goes in: documents and PDFs, photos and screenshots, voice notes, typed notes, forwarded email, onboarding answers, WhatsApp chat exports and social data exports. The system types each capture (People, Documents, Facts, Events, Episodes and subtypes beneath them), logs it in an append-only audit trail, and stores it in the UK. Nothing is kept until you confirm it, and duplicates or conflicts come back to you to resolve. In practice the product is a destination you sign in to, not a plug-in bolted onto someone else's chat window. There is a universal input bar for text, voice, attachments, URLs and slash commands, plus Cards that keep frequently needed records live from memory rather than as stale copies. When you ask a question, hybrid retrieval combines typed lookup, vector search, full-text search and facts that know when they were true. Separate Personal and Practice workspaces keep a private life and a professional one apart, and people you invite see only what you allow. It suits professionals who need provenance and consent around what an AI knows (the site cites lawyers and practice owners), as well as individuals juggling family logistics, health appointments, caring responsibilities or founder paperwork. Compared with assistants that forget between sessions, the pitch is continuity; compared with memory features baked into one vendor, the pitch is portability. Memory travels to assistants that support MCP connectors, such as Claude or ChatGPT on a paid plan with developer mode on, and exports as a portable file at any time. The company is explicit about limits: it is not always-on, not a developer API, not a memory model, and it captures only what you ask it to capture. Data is encrypted in transit and at rest, not used to train models, and deletion anonymises your data within 30 days.

How do you use AI Twin?

  1. 1Start your Twin and set the ground rules
    Create your account from the homepage and work through the opening sequence: rules and email, bringing one thing, then naming the Twin. Everything is free for the first 7 days with no card.
    AI Twin — Start your Twin and set the ground rules
  2. 2Bring your first material
    Upload a document, a voice note or a screenshot, or forward an email in. AI Twin proposes where the item belongs and waits for your confirmation before anything is stored.
    AI Twin — Bring your first material
  3. 3Inspect an AI export before you commit
    If you are unsure what a ChatGPT or Claude export contains, run it through the on-device Export Inspector first. Nothing is uploaded during that check.
    AI Twin — Inspect an AI export before you commit
  4. 4Review the trust and data controls
    Read what is captured, where it is stored and how deletion works before you add sensitive material such as medical or client records.
    AI Twin — Review the trust and data controls
  5. 5Connect your assistant
    Point Claude, or ChatGPT on a paid plan with developer mode enabled, at the MCP connector so your memory travels into the assistant instead of being trapped there.
    AI Twin — Connect your assistant
  6. 6Add a practice workspace if needed
    Set up your own Twin first, then add a shared Practice workspace from the workspace switcher so professional material stays scoped away from personal memory.

Pros and cons

Pros

  • Memory lives in AI Twin rather than inside one assistant, so you can swap models without losing contextAI
  • Every capture writes a readable audit entry and waits for your confirmation before it is savedAI
  • Typed entities plus hybrid retrieval give more precise answers than a plain chat historyAI
  • Strong privacy posture: UK storage, encryption, no model training on your content, export any timeAI
  • Separate Personal and Practice workspaces with granular read-only or read/write sharingAI

Cons

  • Extracting value depends on you uploading and curating material, which is ongoing manual effortAI
  • Assistant access is limited to tools supporting MCP connectors, and ChatGPT needs a paid plan with developer modeAI
  • Inference, speech to text and email are processed in the EU, so the UK-only claim covers storage not the full pipelineAI
  • No developer API, so teams wanting to build memory into their own products must look elsewhereAI
  • The 7 day trial is short for judging a product whose value accumulates over weeks of capturesAI

How much does AI Twin cost?

Free trial

Everything is free for the first 7 days with no card required. You add payment details only if you decide to stay, and export and close stay open even after a trial ends.

Learn more

Pricing

Personal is £18 a month and includes one free read-only slot for another person; further people cost £9 a month each, read-only or read/write and changeable. Practice is £43 a month (£18 Personal plus £25 Practice), owner included, with each added person at £9. Annual billing gives two months free. Prices are in pounds sterling with no VAT charged.

Learn more

Integrations

Memory travels out to assistants that support MCP connectors, such as Claude. ChatGPT works on a paid plan with developer mode switched on. Data can come in from email, WhatsApp chat exports and social data exports.

Learn more

Features

Universal input bar for text, voice, attachments, URLs and slash commands; confirmation before any capture is kept; typed memory across People, Documents, Facts, Events and Episodes with versioning and audit; hybrid retrieval mixing typed lookup, vector search, full-text search and time-aware facts; live Cards for frequently used records; separate Personal and Practice workspaces with scoped sharing.

Learn more

Frequently asked questions about AI Twin

  • How much does AI Twin cost?
    Personal is £18 a month and includes one free read-only slot for another person; further people cost £9 a month each, read-only or read/write and changeable. Practice is £43 a month (£18 Personal plus £25 Practice), owner included, with each added person at £9. Annual billing gives two months free. Prices are in pounds sterling with no VAT charged.
  • Does AI Twin offer a free trial?
    Everything is free for the first 7 days with no card required. You add payment details only if you decide to stay, and export and close stay open even after a trial ends.
  • How do you use AI Twin?
    The walkthrough on this page covers 6 steps: 1. Start your Twin and set the ground rules 2. Bring your first material 3. Inspect an AI export before you commit 4. Review the trust and data controls 5. Connect your assistant 6. Add a practice workspace if needed.
  • What platforms does AI Twin support?
    AI Twin is available on Web App.
  • What does AI Twin integrate with?
    Memory travels out to assistants that support MCP connectors, such as Claude. ChatGPT works on a paid plan with developer mode switched on. Data can come in from email, WhatsApp chat exports and social data exports.
  • What are the limitations of AI Twin?
    Extracting value depends on you uploading and curating material, which is ongoing manual effort. Assistant access is limited to tools supporting MCP connectors, and ChatGPT needs a paid plan with developer mode. Inference, speech to text and email are processed in the EU, so the UK-only claim covers storage not the full pipeline.

Status

StatusActive
Views0
Outbound clicks0
Added10/22/2026

Platforms

Web App

Pricing

Free TrialSubscription