Pick a report to begin. Your answers and results stay on this device.Pick up where you left off, or start a new report.
Instagram feed reading
Hand over your posts and see which of ten rooms your feed comes closest to, with the posts behind every reading.
Read my feedFree · About 5 min
INSYour own posts only · Evidence behind every reading
Hi, human.
Hand over posts from your own account and find out whether your feed reads as a gallery, a diary or a plaza. Every reading comes with the posts behind it.
It never reads anyone else's account, and never asks for an account name.
A look inside
Every interpretation in a feed reading carries the posts it rests on. Posts that cut against it are shown too, with what would make it wrong.
SampleInstagram feed reading@sample
Pattern
This feed gives most of its space to new places and landscapes. It reads as treating going somewhere unfamiliar as the thing worth recording.
If the same neighbourhood and the same café kept reappearing, this reading would be wrong.
Tap a post number to see that post. This is a sample, so illustrations stand in for the photos. This is a sample report. It is not anyone's real result; it was written to show how a report is put together.
More about the feed readingMore reports
Six reports in all. Five questionnaire reports, built on published research scales, sit beside the feed reading; each reading shows its scores and how confident it is.
No sign-up. Your answers stay on this device.
The deep reading of your Instagram feed is only in the app. The questionnaire reports and the basic reading are free on the web as well.
Questions
Starting with the parts that may differ from what you expect.
No. There is no way to look an account up by its handle. The feed reading reads only posts from your own account that you hand over yourself: you can pick screenshots of your posts, or choose the archive from Instagram's “Export your information”.
Instagram has no API through which an outside service can fetch the posts of a personal account, and its Terms of Use prohibit automated collection without its permission. Human.ai does not scrape Instagram.
No. It does not accept another person's account or posts, and it does not interpret other people who appear in your photos. Reading a person from their posts happens only when that person asks for it themselves.
If you are curious about someone, invite them to do their own on their own device. Whether to show you the result is theirs to decide.
A questionnaire report is your own answers, scored by fixed rules. The same answers always give the same scores; the answers themselves are a self-description, and they can shift with your mood and circumstances that day. Percentiles are estimates against research samples that answered in English, not Korean norms, and where no comparison data exist no figure comparing you with other people is shown.
The feed reading is not a tool for working out what a person is like. Research finds that personality can be inferred from social-media posts only with low accuracy, with correlations of about 0.2 to 0.4 against personality questionnaires, and how well Human.ai's own feed reading agrees with one has not been measured. That is why counted facts are kept apart from interpretations, and why each interpretation carries its supporting posts and what would make it wrong. If an interpretation does not fit, it may simply be wrong.
No. No report is a diagnosis, and none says that you have or do not have a condition. They are reference material for understanding yourself.
The self-checks (Mood check-in, Focus style report) are for looking at how you are doing now. Whatever the report says, they show where to find help.
If you are worried about your health, talk to a doctor or another qualified professional.
The answers you choose, your drawings, your task records and your finished reports are stored only on the device you are using. There is no account, and the Human.ai server keeps no answer and no report.
In Settings you can export your data as a file, or erase everything on this device in one step. Uninstalling the app or clearing this site's data in your browser removes it too. Human.ai keeps no copy on its server, so there is nothing more to ask us to delete.
If you have asked for an AI interpretation, what was passed on then is deleted by Anthropic, PBC within 30 days of receiving it. Content flagged as a usage-policy violation may be kept for up to 2 years.
AI is used only when you ask for it and agree.
In a questionnaire report it adds a chapter that reads your scores together, and that chapter is marked “AI-generated”. Without AI the report is still complete, written in sentences selected by scoring rules. What is passed on is your scored results, such as scores and type, and never your answer to a single question.
In a feed reading it reads the photos and captions of the posts you selected, sorts them into categories and proposes interpretations. What goes into the report is decided by rules written in code, not by the AI.
What is passed on goes through the Human.ai server to the Claude API of Anthropic, PBC, in the United States. The Human.ai server does not store what it receives. Anthropic, PBC deletes what it receives within 30 days (content flagged as a usage-policy violation: up to 2 years) and does not use it to train AI models.
AI-written text can be wrong, and it has not been reviewed by a human professional.
The self-checks (Mood check-in, Focus style report) use no AI at all, and nothing of theirs is passed on.
Adults aged 19 and over. The first time you start, you are asked for your date of birth once, and the date you enter is not stored.
The questionnaire reports and the basic Instagram reading are free. Only the deep reading of your Instagram feed is sold, inside the app, for ₩13,800: one purchase, one reading. Nothing is sold on the website.
Human.ai is not affiliated with, sponsored by, or endorsed by Instagram or Meta Platforms, Inc.
Human.ai is not affiliated with, sponsored by, or endorsed by Instagram or Meta Platforms, Inc.
