How We Curate — What Earns a Listing on Uptism
Explore the searchable directory, try Stumble, or suggest a find. Every submission requires editorial review before publication. The category hubs and collections are reading guides. There is no email newsletter or promised delivery schedule.
Uptism is hand-curated, not scraped. This page is the trust spine of the whole site: what gets a listing, what gets cut, and the promises we hold ourselves to so a listing here actually means something. If we ever fall short of what is written here, treat it as a bug and tell us.
We are a new, small, independent site. We would rather earn your trust slowly with an honest, curated shortlist than win it fast with a bloated list you cannot trust.
Read the directory framework Read: what earns a listing
Who this is for
- Builders and the builder-curious deciding whether a tool is worth their time.
- Investors, operators, and researchers who want honest signal above directory noise.
- Skeptics and journalists checking whether this is a real editorial project or another link farm.
If you want the short version: keep reading — the "Plain-language summary" below is the whole thing in six lines.
What earns a listing
Every entry has to clear three plain tests before it goes in. We are looking for signal, not volume — a thing a real person could actually use, understand, and be genuinely glad they found.
- Real substance. Something that exists and does what it says. Shipped software over a waitlist, a working project over a press release, a company doing the thing over a pitch deck. We flag what is early, but we do not list vaporware as if it were real.
- Honest framing is possible. We only list what we can describe honestly: what it is, who it is for, and when to use it — plus the trade-offs and limitations. If we cannot write a fair, useful entry without hand-waving, it does not go in.
- Useful to a real person. A builder, an investor, a curious reader — someone specific should be better off for finding it. "Interesting to us" is not enough; it has to be worth a stranger's time.
- It fits the point of view. Uptism is about the future being built — tools, projects, and ideas that make the case for grounded optimism. See the case for optimism for what that means in practice.
What gets cut
Saying no is most of the job. An entry gets rejected — or pulled — when it fails the tests above or trips any of these:
- Vaporware and demo theater. Staged reels, no changelog, no pricing, no way to actually try it. Announced is not the same as shipped.
- Nothing honest to say. If the only accurate write-up is marketing copy, we pass.
- Slop. Thin, auto-generated, or copycat products that add noise rather than signal.
- Overclaiming. Anything that leans on "this will definitely" or cure-all promises we cannot support.
- Pay-to-play requests. Being offered money for a listing is an automatic no, not a maybe. Use the published editorial reading routes and curation framework. No catalog entry, selection result or update cadence is evidenced by this page.
A cut is not a verdict that something is bad. It usually means it is too early to frame honestly, or it is not a fit for what we cover. Things can earn a listing later — you or the maker can use the find self-check and we will take a fresh look.
Our promises
These are the standards you can hold us to. They are the reason a listing here is worth trusting.
Listings are never for sale
No pay-for-placement, ever. Money never buys a spot in the directory, a higher position, or a kinder write-up. Where a link is an affiliate or referral link, it is labeled and it does not change whether or where something is listed. See the full disclosure, and read how directories become link farms — and how we avoid it.
Every entry must be dated
An entry must carry a date and a "last reviewed" stamp. Tools change, projects stall, and prices move, so an entry that has drifted should be updated, flagged, or removed. A date is a promise that we looked recently — not that nothing has changed since.
We say "we curated," not "we tested"
Most entries are curated: chosen on judgment, research, documentation, and reputation. That is different from hands-on testing. We only say we tested or tried something when we genuinely used it, and we keep that label honest. We will not dress curation up as a lab review.
Risks and limits are named, not hidden
Optimism here is grounded, not blind. When something has a real trade-off, a safety caveat, or a limitation, we put it in the entry rather than bury it. For health and biotech we are extra careful: no cure claims, no medical guarantees, and it is not medical advice. Our full standard lives in the editorial policy.
How "stage of the future" is assigned
Alongside its category, most entries get a stage — roughly emerging, building, frontier, or proven — so you can browse by maturity and know how far along something really is. A few things to be clear about:
| Point | What it means |
|---|---|
| It is human judgment | An editorial read, not a score from a formula or an objective metric. |
| It is dated | A stage reflects our read at review time; it can move up or down when we re-review. |
| It is not investment advice | "Frontier" means early and exciting, not "buy." No return or valuation claims. |
| We fight stage inflation | Not everything is "frontier." We would rather under-hype a stage than inflate one. |
How we handle claims we can't verify
Sometimes a maker states a spec, a price, a benchmark, or a capability we cannot independently confirm. We do not guess, and we do not launder a claim into a fact by restating it in our own voice.
- We mark unverified claims as unconfirmed and, where it matters, attribute them to the source ("the maker says…").
- We prefer primary sources — documentation, the maker's own material, and official references — over second-hand summaries.
- If a field is unknown, it stays blank or labeled unconfirmed. We never invent ratings, reviews, testimonials, awards, metrics, prices, or specs to fill a gap.
- If we cannot frame something honestly at all, it does not get listed until we can.
AI-assisted drafting, human-verified fields
We use AI to help draft prose — a first pass at a summary, a tidier sentence. We do not let it decide what earns a listing or invent the facts.
- Structured fields are human-verified. Stage of the future, dates, pricing, license, category, and safety notes are set and checked by a person, not generated and trusted.
- A human makes the call. What gets in, what gets cut, and how a risk is framed is an editorial decision, every time.
- No slop about slop. Using AI to fill a directory with thin auto-generated pages is exactly the failure mode we exist to avoid.
The fuller version of this — sourcing, corrections, and the grounded-optimism standard — is in our editorial policy.
Plain-language summary
- Hand-curated, not scraped. A short, honest shortlist beats a big list you cannot trust.
- To get in, a thing must be real, honestly framable, and useful to a specific person.
- Listings are never for sale. No pay-for-placement, and affiliate links never change the picks.
- Every entry must be dated; we say "curated" unless we truly tested it.
- Claims we cannot verify are marked unconfirmed; we never invent ratings, prices, or specs.
- AI drafts prose; a human verifies the facts and makes every include-or-cut call.
Curious what this looks like in practice?
The best way to judge our curation is to use it. Browse honest, dated entries in the directory — or send us something you think earns a spot.