Frequently Asked Questions
Do I need to enter data manually?
No — manual entry isn't the plan. Describe what you want to add in a sentence — "the Fast & Furious series" — and the AI drafts the records for you to review. Point your camera at a bookshelf and Holigram recognizes what's on it and where it belongs. Or import a spreadsheet saved as CSV (here's how). The one thing it does ask for is your take: when something has a rating, Holigram wants to know what you thought before it saves.
Do I need to be technical?
No. Holigram is made for everyone. The user interface is designed to accommodate users who aren't actively using agents or harnesses. For those who are technical, there are plenty of features that can be leveraged to interact with it agentically.
Is my data encrypted?
In transit and at rest, yes — the way any serious service handles it.
The more useful answer is what that does and doesn't buy you. Holigram can read your Holigram: that is what makes grounded answers, working backups and a real export possible. So the promises worth making here aren't cryptographic ones, they're checkable ones — access for outside assistants is granted per collection and revocable, every read is logged with the collection and row count, and View as AI renders exactly what a connected assistant receives from your granted collections (turn on power-user mode in Settings to open it). The detail is here.
What happens to my data when an AI reads it?
Only what it can reach — the collections you opened, plus the facts about you that you haven't locked — and only for that question. It never trains anyone else's model. Every send is recorded in your egress log — which collection, how many rows, how many tokens, never the content. When you connect an outside assistant, what happens to that slice once it arrives is governed by that provider's terms, which is exactly why grants are per-collection and revocable.
Does Holigram train AI on my data?
No. Your personal data is never sold, it never trains anyone else's AI, and it's never shared without a grant you can see and revoke. When Holigram's assistant answers you, the model provider handles that one request and never trains on it: Anthropic keeps nothing after it answers, Google keeps it for a limited period to watch for abuse, and either may keep what its safety systems flag or the law requires. Holigram's own models train on none of your content today.
If that ever changed, it would be an explicit, off-by-default choice, explained to you in plain language first. One limit worth knowing: once you grant an outside assistant like Claude, ChatGPT, Gemini or Grok a collection, what happens to that slice is governed by that assistant's own terms. The Privacy Policy has the detail.
Can ChatGPT, Claude, Gemini and Grok share memory?
Not with each other: each assistant's own memory stays with the company that runs it. What they can share is Holigram. Connect more than one, and each reads the collections you open to it, plus what Holigram remembers about you, which you can open, edit, and delete. Every read is logged, and you can revoke any connection at any time.
When you connect each one, you choose what it may change: Read only, so it reads and never adds anything; Suggest, the default, so anything it adds waits in your approval queue until you approve it; or Save directly, so it saves entries and facts, and sets up new collections, without waiting, each one logged with an alert that lets you undo it. At every level it can't change or delete what's already in your Holigram. The free plan keeps one connection at a time, so connecting more than one at once takes Premium, which early-access testers already have. How to connect them.
Can I get my data out?
Yes, and it's free on every plan: every collection and entry you keep, as JSON, in one tap. No support ticket, no waiting period.
How much does it cost?
Holigram has a permanent free plan, not a trial: 15 collections, 100 documents, one connected assistant and everyday AI use. Paid plans will cover what costs real money to run, like heavier AI use and photo storage. Their prices aren't published yet, and nothing is purchasable during early access, where testers have the full product.
Prices will be on the Plans page before anyone can be charged, and keeping, exporting and deleting your data stay free on every plan. The split is here.
Is there an Android app?
Not a native one. Holigram has an iPhone app on the App Store, and on Android the web app does the job: it runs in your browser, installs to your home screen, and sends notifications. It's the same Holigram on every device, so what you add on your phone is waiting on your computer, and the other way round.
How do I keep track of the wines I've tried?
Add the ready-made Wine collection. Each bottle has room for its producer, varietal, region, vintage, your rating, what you paid, how many you have and where they sit. Add a bottle by describing it in a sentence, then ask, "What was the Malbec I loved?" and the answer comes from your own entries, not a web average.
Drink From and Drink By years ride along too, so "What's ready to open?" has an answer.
How do I keep track of my family's clothing sizes?
Use the ready-made Brand Sizing collection: one entry per brand and kind of clothing, with the size and how it fits. For a family, add a field for whose size it is; adding your own fields is free. Then ask, "Which size does Maya wear in Levi's jeans?" before you order, and the answer comes from what you kept.
Share the collection with your household, and everyone in it can add and check sizes too.
Who can see what I share?
Only the people or groups you shared with, only the collection you shared. A friend sees the one collection you co-authored with them, not your Holigram. A Prism shows only the two paired collections, only the columns you both keep, and never ratings. Every one of those can be ended by either side.