frequently asked questions

Plain answers about what Anamnesis is, how to set it up, and what happens to your data. Anything this page doesn't cover, ask us directly on the contact page.

01What is Anamnesis?

Anamnesis is a memory for your AI tools. Normally, every chat with an AI assistant starts blank: it has no idea what you told it yesterday, so you re-explain your project, your preferences, and your decisions over and over. Anamnesis stores that knowledge for you and hands the relevant parts back to your assistant at the start of every conversation.

02How does it work, in plain terms?

Think of it as note-taking that runs itself. As you work, your sessions are saved as raw records ("episodes"). Anamnesis then distills each one into a short, durable note ("echo"): the decision you made, the fact worth keeping, the preference you stated. Over time, related notes consolidate into long-term knowledge ("engrams"). When you start a new conversation, your assistant asks Anamnesis what matters right now and gets back just those notes instead of a wall of old transcripts.

The names come from real memory science, and the stages mirror how human memory encodes and consolidates experience.

03How do I get started?

1. Create an account. Free during alpha, no credit card.
2. Copy your connector URL and API key from the dashboard.
3. Paste the connector into your AI tool. Each tool's settings has a place for MCP connectors or integrations; the dashboard shows per-tool instructions.

Setup is one URL, once. From then on the memory follows you into every session.

04Which AI tools does it work with?

Claude (Claude Code, Claude Desktop, claude.ai, Cowork), ChatGPT, Codex CLI, and Gemini CLI. One account, one memory, shared across all of them: something you decided in Claude Code is known to ChatGPT the next time you open it.

One honest distinction: fully automatic session capture runs in Claude Code today. The other tools read your memory into every chat, and save new memories when you ask them to ("remember this") or when the assistant judges something worth keeping.

05Can't I just build this myself with RAG?

The retrieval part, yes. Embedding your notes and searching them is a weekend project, and for a small collection used from one tool it works fine. The hard parts are the ones around it. Somebody has to write the notes, so capture runs automatically. A wrong or malicious line has to be kept from becoming part of what your AI believes, so everything is screened on the way in. Decisions get retired when they change instead of lingering next to their replacements. And ChatGPT or claude.ai will never run your local scripts, but they will connect to Anamnesis. The benchmarks put numbers on each of these.

06Is my data private?

Yes, and we've written down exactly what that means. Your memory is encrypted with a key derived from your own credentials, we don't train models on it, we don't sell it, and the site runs zero third-party trackers. The full technical picture, including the threat model and what we don't yet claim, is on the security page; the legal version is in the privacy policy.

07What does it cost?

Nothing during the alpha. The long-term business model is a subscription. There is no ad tier and no data-sale tier, at any price, ever.

08Can I see and manage what it remembers?

Yes. The dashboard lists every memory you have, searchable and browsable, and any one of them can be deleted. Deletion is permanent: the encryption key is destroyed, so even our rolling backups become unreadable.

09What if it remembers something wrong or outdated?

Tell your assistant the correction, or delete the memory from the dashboard. Anamnesis also tracks facts over time on its own: when a newer memory contradicts an older one, the newer dated version wins, and stale beliefs are retired rather than repeated back to you.

10Can I take my data and leave?

Anytime. Export downloads your whole corpus as plain JSON, with no proprietary format in the way. Delete your account and the data is gone for good.

11Something isn't working. Where do I get help?

Tell your connected assistant to "report an issue with Anamnesis" and it will file a structured report for you, or reach us directly through the contact page or at hello@smtry.ai. A real person reads every report.