Most storefronts are good at showing you what is already moving. That is useful when you want the game everyone is discussing. It is less useful when you want something odd, small, old, or made for a very specific kind of player.
Finding hidden indie games takes a different habit. You have to search sideways rather than stare harder at the front page. The goal is not to beat an algorithm or collect hundreds of wishlists. It is to build a short trail from one game you love to the people, tags, festivals, and communities around it.
This guide lays out a practical way to do that without turning discovery into a second job.
Start with a feeling, not a genre
Genre labels are broad. "Roguelike" can mean a quick deckbuilder, a punishing action game, or a six-hour strategy session. "Cozy" can cover farming, decorating, puzzles, and stories about grief. A genre search often gives you a technically correct pile that still feels wrong.
Start by describing the experience you want. Maybe you want a game that:
- fits into one quiet evening
- lets two friends solve things together without rushing
- feels strange but not cruel
- has a town you can learn slowly
- uses one unusual mechanic and stops before wearing it out
Those phrases are better filters than "indie" on its own. Turn them into combinations of tags, then test several versions. "Short atmospheric puzzle" will produce a different shelf from "narrative puzzle under five hours." The words do not need to be elegant. They need to describe the night you want to have.
Follow one good game outward
The best discovery tool is often a game you already trust. Look past its store page and follow the links around it.
Check the developer's earlier work. Small studios often revisit an idea across several games, and solo developers may have game-jam projects that show the same voice in rougher form. Look at the publisher too. A focused publisher can act like a record label: you may not love every release, but the catalogue tells you what kind of risks it finds interesting.
Then look for the people recommending that game. A critic, curator, streamer, or forum regular who noticed one obscure title you loved may have ten more waiting nearby. Taste is more useful than scale here. The person with a small, consistent beat can be a better guide than a general list built to cover everything.
Steam's own discovery page describes Curators as individuals, organizations, and groups that publish public recommendations. It also mixes friend activity, personalized suggestions, tags, and a Discovery Queue into the storefront. Those tools work better when you treat them as leads rather than verdicts. Open the page. Read the negative reviews. Watch ten minutes of unedited play. Decide for yourself.
Use demos as evidence
Trailers tell you what a game wants to be. A demo tells you what it feels like in your hands.
That difference matters most with small games, where the hook may be a movement system, a peculiar interface, or a rhythm that screenshots cannot explain. Give a demo enough time to reach its central idea, but do not force yourself through it out of politeness. Friction is useful information.
Steam Next Fest is built around this kind of sampling. Valve describes the event as a multi-day celebration where players can try demos, talk with developers, watch streams, and learn about upcoming games. The October 2026 edition runs from October 19 to October 26. That week will be noisy, so going in with two or three experience-based searches is more productive than downloading whatever sits at the top of the most-played list.
A good festival routine is simple:
- Pick two moods or mechanics before browsing.
- Install no more than five demos at once.
- Play each one before adding another batch.
- Keep only the games you would genuinely make time for.
The limit is important. A folder of 40 untouched demos tells you nothing except that browsing is easier than playing.
Browse places where rough edges are allowed
Large stores reward presentation because a capsule image has to win a click in seconds. Some of the most interesting games are not ready for that contest, and some never want to enter it.
Itch.io is useful because its shelves include free games, browser games, demos, game-jam entries, tools, soundtracks, and tiny paid releases in the same ecosystem. Its browse pages expose tags such as horror, multiplayer, visual novel, simulation, roguelike, and game engines, while the site also surfaces jams and developer logs. That messiness is a feature. You can move from a finished release to a weekend experiment and see where an idea began.
Game jams are particularly good for discovery because constraints push developers toward clear, compact ideas. The result may be unbalanced or unfinished. It may also do one thing you have never seen before. If a jam game catches your attention, follow the creator. The polished commercial game often arrives much later.
Do not judge these spaces by the standards of a major release. Read the page, check platform support, scan comments for technical problems, and decide how much uncertainty you are comfortable with. Curiosity does not require abandoning basic caution.
Find a human filter with a narrow beat
Broad "best indie games" lists tend to converge. Narrow coverage is where the useful surprises live.
Look for a writer who covers interactive fiction, a channel devoted to automation games, a newsletter about short horror, or a community that cares about local co-op. The narrower the beat, the easier it is to understand the recommender's taste and disagreements. You are not looking for someone who is always right. You are looking for someone predictably wrong in ways you can account for.
Pay attention to the reasons behind a recommendation. "Charming" tells you very little. "The dialogue changes when you repeat a failed day" gives you something to test. Specific criticism travels better because you can compare it with your own preferences.
Friends are useful for the same reason. Ask what surprised them, not what scored highest. Ask which game they still think about, which mechanic they have tried to explain badly three times, or which release they wish had found a larger audience. Those questions produce stories rather than rankings.
Read the quiet signals on a store page
A small game's page contains clues that have little to do with its review average.
Check whether the description can explain the central action in plain language. Watch for a demo, an active update history, readable accessibility information, and developer replies that answer questions directly. Look at the dates on reviews. A game that had technical problems at launch may be in a different state now, while an old positive consensus may no longer fit an abandoned multiplayer game.
Negative reviews deserve attention, but sort the complaints. "Too short" could be a recommendation if you want a focused evening. "No map" may be central to the design. Repeated reports of crashes are different. The useful question is not whether complaints exist. It is whether the same issue appears across unrelated players and whether that issue matters to you.
Review counts can signal visibility, not quality. A low count should make you gather more evidence, not dismiss the game. Try the demo, find footage without commentary, or read a developer post. Hidden gems are hidden partly because fewer people have supplied the reassuring signals that make a purchase feel safe.
Keep a small discovery shortlist
Discovery becomes frustrating when every interesting game goes into one enormous wishlist. Separate "I noticed this" from "I want to play this soon."
Keep a shortlist of five to ten games. For each one, write a sentence about why it caught you. "Two-player mystery with no combat" is more useful three months later than a title you no longer recognize. When the list is full, play, remove, or replace something before adding more.
Perthro can hold games you plan to play in a wishlist, let you reorder that list, and keep a "next up" view. You can also make custom lists for narrower trails, such as festival demos, one-evening games, or recommendations from a particular friend. Perthro is free to download on the App Store and requires iOS 17.0 or later.
The point is to preserve context. A game recommendation usually arrives attached to a person, mood, or moment. Without that context, even a great find can turn into another unfamiliar cover in a long row.
Share discoveries while they are still small
A hidden game stops being hidden one player at a time. If something works for you, say why.
A useful recommendation does not need to be long. Name the specific thing the game does, mention who might enjoy it, and be honest about the rough edges. "A two-hour detective game where taking notes is the main mechanic, but the text is tiny on a television" helps more than "underrated masterpiece."
Perthro lets players rate games on a five-star scale, write reviews of any length, follow friends, and react or reply in a shared feed. That gives a recommendation some context: it comes from a person whose taste you can learn over time. If you want to compare notes beyond the app, visit the community. For account or app help, use support.
A better discovery loop
You do not need a perfect recommendation engine. You need a loop that keeps returning you to play:
- Describe the feeling or mechanic you want.
- Search tags, festivals, jams, and focused critics.
- Test the idea with a demo or unedited footage.
- Save a small number with a note about why.
- Play one.
- Share what you learned with someone else.
That final step matters. Storefronts can show similarity, popularity, and momentum. A person can tell you that a strange little game arrived at exactly the right time. That is usually where the best discoveries begin.