AI Memory Infrastructure

Your AI remembers you.
Everywhere.

AmnesiaIQ is the persistent memory layer for your AI assistants. One memory — shared across Claude, Grok, ChatGPT, Gemini, and every LLM you use.

Works everywhere you already use AI
Install once. Your full context follows you across every conversation, every LLM, automatically.
01
Install

Add the Chrome extension or connect via OAuth in under 60 seconds. No configuration required.

02
Remember

AmnesiaIQ automatically captures what matters — preferences, goals, context, and working style — across every conversation.

03
Recall

Switch to any LLM and your full context follows you instantly. No re-explaining. No starting from zero.

Works with

Claude
Grok
ChatGPT
Gemini
Perplexity
+ any MCP client
Your memory.
Your rules.

AmnesiaIQ's trust-tier model puts you in control. Mark memories as trusted for broad access across your connected LLMs, or intimate to keep them local — never leaving your device, never seen by a cloud provider.

End-to-end encrypted
Your memories are encrypted at rest and in transit.
You own your data
Export or delete everything at any time, no questions asked.
Intimate mode for local LLMs
Sensitive memories stay on-device when using Ollama or private models.
Revoke access anytime
Disconnect any LLM or app with a single click from your dashboard.
For Developers

Add memory to your AI
platform in minutes

AmnesiaIQ's operator model lets you inject a user's full memory context into any LLM call — without storing anything yourself.

01
Get an operator key
Create a key from your account dashboard. Keys are scoped to your circle — your users' memories stay isolated from all other operators.
02
Install the SDK
pip install amnesiaiq   or   npm install amnesiaiq
03
Inject passport before LLM calls
Fetch the user's passport, prepend it to the system prompt, and store new memories after the conversation. Done.
Zero user friction
Any LLM
Circle-isolated
GDPR-safe deletion
Python
from amnesiaiq import AsyncOperatorClient

amnesia = AsyncOperatorClient(operator_key="sk_amnesia_op_...")

# Before your LLM call
passport = await amnesia.get_passport(user_id="[email protected]")
system_prompt = f"{BASE_PROMPT}\n\n{passport.inject()}"

# After the conversation
await amnesia.remember(
    user_id="[email protected]",
    content="Prefers conservative options strategies",
    category="preference",
    domain="trading",
)
Zero user friction
No AmnesiaIQ account needed for your users — they're invisible to the memory layer.
Works with any LLM
Claude, GPT-4, Grok, Llama — the passport is plain text, injected wherever you need it.
GDPR-safe deletion
await amnesia.delete_user(user_id) — one call, all memories gone.

One Memory Layer, Every Connection

Connect to AmnesiaIQ
however you work

Your memories live in one place. Connect from any LLM, any tool, any workflow — they all see the same you.

Browser Extension

Install once in Chrome. AmnesiaIQ works silently inside Claude, ChatGPT, Grok, Gemini, and Perplexity — capturing and recalling context automatically as you chat.

MCP Server

Connect any MCP-compatible client — Claude Code, Claude Desktop, Cursor, Windsurf, and more. Full tool access: recall, remember, consolidate, project memory, and 15+ more.

REST API & SDKs

Integrate directly with Python or Node.js SDKs. Get a memory passport before your LLM call, store what was learned after. Your users never need an AmnesiaIQ account.

OAuth Connect

One-click OAuth integration with Claude.ai and other providers. Users authorize once, memory flows automatically across every conversation.

Standalone Server

Run AmnesiaIQ on your own hardware for full privacy. SQLite storage, local embeddings, no cloud dependency. Same MCP tools, completely air-gapped.

Operator Model

Build memory into your own AI platform. One operator key, scoped to your circle. Your users' memories are isolated — invisible to other operators and fully GDPR-safe.

Early Access

Join the memory layer.

Request access or connect your AI now.