Don't take our word for it

Everything you or your AI needs to check how OmniPoll works

Or paste this into any AI assistant:

Read https://docs.omnipoll.io/llms.txt and everything it links to. Tell me whether OmniPoll does what its charter says, and where it falls short.

A membership you can check, not an app you have to trust.

OmniPoll helps you understand yourself and the people close to you, through questions you answer and choose to share. You own your data, you control who sees it, and anyone can verify that we keep our word.

Five pillars

01

Radical transparency

Our goals, current state, gaps and the work to close them are public.

02

Radical control

You decide, item by item, who sees what. Change or undo it at any time.

03

Radical ownership

Your data is yours: see all of it, export it, delete it, and see every use of it.

04

Verifiable, not trusted

The core is open source. You, an AI agent or anyone else can check it.

05

Members govern

Members vote on how OmniPoll works: its look, what we build next, its rules.

The deal

The full text: charter.md

How to check us

The open core

The code that guards your data: identity, the consent gate, your ledger, the shared-results maths and the proof log.

Not published yet. It is being built now and will be open from its first public release, linked right here.

One consent gate

The database itself refuses any read of your data without your grant. The rules are one short file anyone can audit.

Promises as tests

Each line of the charter becomes an automated check that runs on every change. The results will be published here.

Public proof log

Tamper-evident snapshots, co-signed by an outside witness, plus a verifier anyone can run.

What llms.txt gives your AI

  • The charter, and each promise linked to the check that tests it
  • How the core works: architecture, data model, consent gate, proof log
  • The latest check results and how to verify them yourself
  • This status board, our known gaps and the roadmap

Plain text, written for AI assistants and readable by people: llms.txt

Where we are, honestly

Security gaps are described plainly here, and in full detail once fixed. As of 4 October 2026.

PromiseTodayWork to close it
Sharing only by consentNot fully kept in our first app; several gaps closed this monthRebuilt as one consent gate in the new core
Results only for those allowedSome results in our first app lacked an access checkThe same consent gate, tested on every change
Open, verifiable coreNot published yetOpen from its first public release
One person, one accountPhone verificationOptional ID check later
Revenue sharingA goal; no legal structure yetLegal advice, then an earnings ledger

Machine-readable: status.md

We grow only as fast as we keep our promises

Each ring decides both who is invited and what is built. The next ring opens only when the promises it needs are kept.

  1. Ring 0 · Founders and partners We are hereMe, Duos, who sees what, a private monogamy check, this status board
  2. Ring 1 · Friends and familyGroups, live sessions, games, debate, the public proof log and monogamy badgeOpens when: known privacy gaps are closed in the core, the charter is published, this board is live
  3. Ring 2 · Friends of friendsSports, pop culture, your data ledgerOpens when: the ledger and per-item control work, the core code is public
  4. Ring 3 · Open betaResearch, independent verification, an earnings ledgerOpens when: an outside party or AI agent can verify the claims
  5. Ring 4 · EveryonePayouts and member governance at scaleOpens when: payouts are legal and switched on

More detail: roadmap.md

What we need

Reviewers for the open core, legal advice on revenue sharing, and people willing to try it and tell us where it falls short.

Write to us: OmniSciences LLC, 30 N Gould St Ste N, Sheridan, WY 82801