Our first months, in the open.
Maitreya is about eight months old. This is not an annual report — it is a founding report: small, honest, receipts-first. Real numbers, including the zeros.
We start with what we got wrong.
What I learned is that child-safety-first and trust-first internet design — starting by prioritizing this — is the right choice. It is not an afterthought. Why? Recent news from Meta showed that when it was an afterthought, patching isn't enough.
And despite how hard the work is — speaking amidst the flood in Nepal, and the sadness of climate overtakes me — I wonder how we can build upon what we knew already to create a better world for our future generation, and this part of the work I am doing is pivotal.
What we have in place — you can dig into our codebase — is a detection system and backend for detecting, flagging, and responding to materials on active abuse and suicidal ideation, and responding protocols are actively being designed. This governance makes and breaks what we do here at Maitreya, Inc., and we take it more seriously than anything else.
I am the Chief Executive Officer and Haleigh is our Chief Brand & Ethics Officer. These days I design, scope, and operate, and Haleigh provides oversight. Our operating model is AI-native: I direct AI agents with my judgment and taste, and I direct them away from AI slop. I orchestrate the agents.
We are small on purpose, because it lets us move fast. Trust-first infrastructure means trust is considered at every single level, including at the cofounder and founding-team level — we have to be mission-aligned. I expect my cofounder, and everyone who comes after, to be ready to use technology to their advantage, and I am personally launching the Operators Room to teach people how to do exactly that.
Trust is grown, not granted. That includes yours in us — so here is everything, written down.
We are a public benefit corporation. That is a legal obligation, not a slogan.
Our charter binds us to a stated public benefit: trust infrastructure for the people the internet left behind. These commitments are fixed — they do not change with growth pressure.
One two-minute reel about being groomed. This is what came back.
In the comments and messages under that reel, survivors disclosed things they had never said out loud. A comment section is not built for that. That gap — between what people carry and where they can put it — is the demand we are building for.
These are platform-reported figures for a single breakout reel, read directly from the account dashboard. They measure attention, not need — we cite them as a signal, not a proof.
The infrastructure is operational. The honest number is zero.
Capital-efficient by conviction, not just necessity.
A company that runs lean can afford to keep its promises. We are not publishing a P&L or vendor list at this size — we are publishing the number that matters.
Promises you can verify against the next report.
As a public benefit corporation, we state plainly where we fell short of our stated benefit this period: we have not yet delivered a sustained match between the people our platform exists to connect. Near-matches have happened; none has sustained yet — visitor volume is still low and we are early in development. The infrastructure is live; the outcome is not. We expect to report against this line, honestly, in every report until it changes — and after.
What the community response was — and what it was not.
A reel about grooming reached more than 528,000 people. About 790 people left public comments. Roughly 130 shared their own experience of childhood abuse, often naming the trusted adult who did it — a family member, a babysitter, someone at school. About 26 asked to be part of what we're building.
We are not quoting any of them here.
No one who commented agreed to have their words placed in a company document. They were speaking to a moment, not consenting to a founding report. Many disclosed abuse and named people close to them. Turning that into pull-quotes — even anonymized, even to make a good-faith point — would take something given for one purpose and spend it for another. We won't. A consent-first process is running separately to decide what, if anything, we can ever responsibly use.
And what this is not: reach and comments are engagement, not evidence our product works. By our own count, we have produced zero sustained matches to date. The response tells us the problem is real and widely felt. It tells us nothing about whether our solution functions — and we won't let one stand in for the other.
To the people who reached out: we heard you, and we haven't used you. If there is ever a consent-first way to bring your experience into this work on your terms, that invitation is yours to accept or decline.
Where every number on this page comes from.
How we gathered every number in this report — one entry per figure: what it measures, where it came from, and what it does not tell you.
Source: Instagram native Insights for the grooming reel, as reported by the account (lifetime figures as of this report). 'Reached' counts accounts that saw the reel; 'views' counts plays. Neither counts unique individuals across platforms, nor retention, nor whether anyone was helped.
Source: billing consoles, read August 29, 2026. Claude (Anthropic API): trailing-12-month average of paid invoices, about $140. Supabase: current Pro plan, $25. Resend: current email plan, $20. Cloudflare: founder estimate, $5. Excludes the founder's own time. An estimate, labeled as such.
Source: internal platform records as of August 29, 2026. Counts connections that persisted past first contact. Near-matches have occurred; none has sustained.
Corrections: if we got a number wrong, we will amend this page and log the change here, dated. Write to hello@maitreya.company.