Really well put — the psychological difference between “claim” and “submit” is real and I hadn’t framed it that way before.

Worth mentioning that the dev upload feature wasn’t something I designed in isolation. It came directly from feedback from indie devs who asked for it. So while the claim model solves one population, there’s another population that actively wants to own and manage their listing from day one. Both are real.

A few honest constraints on the itch.io crawling path: itch.io’s API is developer-only (your own sales, download keys) there’s no public catalog API to browse or ingest games the way IGDB provides for mainstream titles. So coverage would require a scraper, which is technically possible but would likely violate their Terms of Service, and is the kind of thing that gets you banned or ignored if you ever try to reach out for a partnership later. That’s also why direct dev uploads are cleaner for now; I control the relationship, and devs get a dedicated section in Discover. And a crawler would undercut exactly what makes the upload flow valuable: a dev who submits their game gets immediate visibility, their own dedicated page, and a section built specifically for dev-submitted games. That’s something a crawler can never replicate.

On the claim model specifically, you’re right that it removes the self-promotion barrier, but it does hand control of the listing to someone who never asked for it. Platforms like Steam and GOG go the opposite direction and require devs to own their page from day one. There’s a reason for that: devs who care about how their game is presented want to control it, and an auto-generated page can misrepresent them before they even know it exists. The opt-out you mentioned is the minimum but the first impression problem is real.

On the claim model from an engineering perspective, I love the idea and “your game is already listed, do you want to claim it” is a much better pitch than a submission form. The challenge is verification: how does someone prove they own the game? Without that, a bad actor creates an account and starts claiming and editing games they don’t own. Discogs works because music ownership is verifiable through external signals like label credits, artist profiles, release history. A game dev claiming their itch.io page needs a similar anchor, probably proving ownership of the itch.io or Steam URL. That’s solvable but not trivial. IGDB is actually a closer comparison than Discogs here since it’s a community-maintained database where anyone can edit, and it works because games have enough external reference points to catch bad edits. The difference is IGDB has a large active community to self-correct and is not just indie oriented. At this scale, every claim needs manual review, which brings back the bottleneck the claim model was supposed to remove. Worth building toward but the verification layer has to come before the claim flow, not after.

On the psychological barrier, you’re right that claiming feels different from submitting. But I’d push back slightly: some friction isn’t always bad. A dev who puts zero effort into visibility is also likely to put zero effort into responding to players, updating the game, or engaging with their community. The platform works better when devs are at least a little invested. The goal isn’t to remove all effort, it’s to make the effort feel worthwhile. If a dev sees that listing their game here drives real traffic and real players, the motivation follows. That’s the bet. And players notice. Gamers genuinely respond to devs who show up who post updates, respond to feedback, engage with their community. Some of the most beloved indie games probably aren’t the most polished ones, they’re the ones where the dev was visibly present. That engagement is part of the product, and it starts with being findable.

The deeper issue with “never market” devs is motivation, not friction. Even a claim flow requires them to care enough to act. Some won’t, and that’s genuinely hard to solve without something pulling them in: visibility, downloads, a reason to show up. That’s what the dedicated dev-uploaded section is trying to build toward: if a dev sees real traffic coming from a listing, that’s the motivation to engage.

On Bombercup, I don’t have a crawler yet so I can’t pull it in automatically, but I’d genuinely love to have it on the platform. It sounds like exactly the kind of game that deserves more eyes. A few options: I can add it manually as admin, or you can upload it yourself through the dev upload flow. It has its own dedicated section for dev-submitted games. Either way works, happy to do whichever is easier for you.

On your cold-start challenge, it’s a fair test and I want to answer it honestly. Multiple moods are supported simultaneously so chaotic, competitive, and nostalgic working together isn’t a problem. On comparative search: it partially works. Searching “atmospheric horror” or “action arena” works because the system searches across tags, genres, themes and descriptions so descriptive language maps reasonably well. Where it breaks down is truly comparative language: “games like Bomberman but faster” or “something like this but shorter and more relaxing” — the system has no way to parse that kind of relational intent yet. That’s a bigger challenge and an unsolved one. On mood inference: moods are currently generated at ingest time and in the case of dev uploads the dev defines them directly. The longer term direction is player-driven: after playing a game, players log how it made them feel, which over time builds a much richer and more accurate signal than anything inferred from text alone. That’s harder to game, harder to hallucinate, and closer to how people actually talk about games. Not there yet, but that’s where i am thinking on going.

I genuinely appreciate the thoughtfulness here. This kind of feedback is rare and it does shape how I think about what to build next. That said, every feature decision sits at the intersection of what different people want, and those people don’t always agree. The upload flow came from devs who asked for exactly that. The claim model is a valid alternative that solves a different problem. Both represent real perspectives, and neither one alone defines the right direction. I’m one person navigating a lot of competing needs, and the honest answer is that no single conversation settles these questions but conversations like this one move them forward. So thank you.

First of all, thank you for your response I really appreciate it.

You’re raising something real, and I’d rather engage with it honestly than defend a pitch.

A clarification first. I don’t pull from itch.io. The catalog comes from IGDB, which covers commercially released games that went through some baseline quality filter just by existing on those platforms. The 1.2M itch.io problem is real, but it’s not the problem I’m solving right now and honestly, dumping 1.2M unvetted games into the catalog would break the thing I’m trying to build. Quality over volume is a deliberate call, not a limitation.

About the curation. You’re right that tags alone don’t work. That’s actually why the platform combines multiple layers instead of relying on a single signal: mood (“atmospheric”, “cozy”, “tense”), genres, themes, completion time, platform, and rating filters. Each layer narrows the search further. Someone looking for a short, atmospheric puzzle game under 5 hours on PC with a high rating gets a very different and much smaller list than a genre tag alone would produce. It’s imperfect, but it’s a fundamentally different bet than tag matching.

On the review point. You’re describing exactly what the platform should grow toward, and honestly what I’m already working toward. A number out of 5 tells you consensus, not fit. The next feature I’m building lets players log not just a rating but how a game made them feel and refine or add tags from their own experience. The mood and vibe signals are an attempt at that: matching on what a game feels like rather than how it scored.

The data integrity question is real: Community tagging can be gamed. But aggregation handles most of it: one bad actor’s vote is noise, a pattern across many players is signal. IGDB solved the same problem with community edits, moderation, and trust scores. The mechanism doesn’t need to be perfect, it needs to be resistant enough that manipulation doesn’t pay off.

One more thing worth naming directly. “hidden gem” is subjective by design. A hidden gem isn’t a game that’s objectively great and undiscovered it’s actually a game that’s right for you that you haven’t found yet. Someone who loves cozy farming games and someone who loves brutal roguelikes are looking for completely different hidden gems. The platform doesn’t curate taste, it narrows the search based on yours.

That’s the actual goal that if you finished a game you loved and want something similar, you shouldn’t have to spend an hour reading reviews, watching trailers, and cross-referencing completion times. You already know what you liked and the platform should use that to get you to the next one faster. The closer I get to a real player profile, what you played, what resonated and what didn’t the more precise that match gets.

What also makes this different from every other IGDB-powered discovery site is that developers can actually participate. They can submit their game, own their listing, and reach players who are actively looking not just browsing a database someone else built. Any dev, including itch.io devs, can go through the upload flow. I just don’t want to pull 1.2M games automatically and call it discovery.

On the “no ratings” point, that’s one of the harder problems to solve honestly, and I’ll be upfront about it. I’m building this solo, which means every design decision, what to approve, what to reject, how to weight signals falls on one person. An approval step helps filter obvious noise, but it creates a new problem: I become the sole judge of what’s worth listing, and that doesn’t scale and isn’t fair. A genuinely good game with zero ratings because the dev had no marketing budget looks identical to a bad game with zero ratings on paper. I don’t have a clean answer to that yet. What I do know is that volume alone isn’t the goal a smaller catalog where every game has a real chance of being found is more useful than 1.2M entries where 99% are invisible anyway. The right mechanism probably involves community input, dev engagement signals, and time but it needs more thinking, and this kind of feedback is exactly what shapes those decisions.

Since you’ve thought about this just genuinely curious: if you were setting the rejection criteria for dev-submitted games, what would you use? Not “is the game good” because that’s subjective, but what signals would tell you a submission is worth listing versus noise? I’m still working that out and I’d rather not get it wrong.

Lastly none of this is black and white. There are real tradeoffs between business logic, engineering constraints, design, and what most players actually want and as a solo developer I’m navigating all of those simultaneously. This is a work in progress. The direction is clear even if every piece isn’t built yet. Where I genuinely agree: without a richer player profile, the recommendations are educated guesses at best. That cross-reference layer is the hard part, and I’m not there yet. But that’s exactly where I’m going.