For publishers bringing games to players in China

Know which games will land in China
before you sign them.

We forecast how a pitch will do with players in China from its text, screenshots, languages and price, before a Steam page exists.

Ranked with forecasts within 24 hours.
Free, under NDA · the pitches you are deciding on now
91%
of the time we pick the future hit over the future flop
5×
more hits when you sign the pitch we rank first, not a random one
30,849
real launches it is checked on
Proof

We scored 464 new games before launch

← lowest chance from ushighest chance from us →
became a hit: 50+ reviews from players in China in the first monthdid nothover a bar to see the game
Recent games

What we said before launch, what happened

Check a game you know

Pick a game you know. Maybe one you signed.

Pick a released game
→
See the chances we'd have given its pitch before launch
→
See the 20 games we scored most like it, and what they got
Your pick

Ten pitches in one genre. You sign one.

100 publishers
→
each gets ten pitches in the same genre
→
each signs one
→
how many end up with a hit?
Each signs one at random
13 hits out of 100
4×
Each signs the one we rank first
53 hits out of 100

Hit: 50+ reviews from players in China in the first month. Counted on 30,849 Steam games released in 2023–2026, each scored by a version of the model that did not see it.

How it works for you

Every pitch scored. The best on your desk.

Pitches come inA deck, or one paragraph with 4 images or references, is enough.
→
Scored in minutesChance of a hit with players in China, a range, place among real launches.
→
The five worth your timeEvery Friday, in writing.
Pitches reach us three ways
you forward the emailwe connect by APIstudios fill in our form
The offer

30-Day Pilot

Run it on your real pitch inbox for a month. Then decide.

Start with 10 free pitches →

What you get

Every pitch rankedOn one board, by its chance with players in China
A forecast with a rangeFirst-month reviews from players in China, low to high
China vs. other marketsWhether the game is stronger there than elsewhere
Closest games that did wellSix comparable launches and their numbers
A weekly shortlistThe five worth your time
One private linkNo sign-up, the whole team opens the same board

Terms

Start free10 of your pitches, ranked with forecasts within 24 hours
Pay if it was worth itA small refundable deposit to start, the rest on day 30
Under NDAYour pitches stay private and are deleted on request
Only 3 new publishers a monthWe write every shortlist ourselves, so places are limited
How it decides

Math on 30,849 Steam launches.

1We collected the marketEvery game released on Steam in 2023–2026: its page in English and Chinese, screenshots, price in USD and CNY, and reviews from players in China in the first month.
→
2Turned each game into numbersGenre, languages and Chinese localization, price, how it looks, how long it plays: hundreds of measurements per game.
→
3Taught a model what hits look likeA statistical model (gradient boosting and regression) learned which combinations ended up as hits. Then we checked it on games it had never seen.
Questions

What publishers ask us first

Will our pitches stay private?

Yes. We sign your NDA or ours before the first pitch. Pitches are used only to score them for you, and we delete them on request. We do not make or publish games.

Isn't this just an AI guessing?

No AI guesses the number. A statistical model built on 30,849 released Steam games makes the forecast, and it was checked on 930 games released in September 2026 that it had never seen. The charts above are that check, game by game.

Where is it wrong?

Some games are hard to call. Games from Chinese studios are almost all strong with players in China, and ranking them against each other is harder than telling a hit from a flop. A cheap global viral hit without a Chinese store page can surprise it, and so can a game that takes off through one streamer moment after launch. By genre, rank correlation on the September 2026 releases: roguelikes 0.80, co-op 0.78, management 0.77, platformers 0.74, idle 0.74, simulation 0.72.

We already run wishlist tests.

Good, the two work together. A wishlist test needs a Steam page, art and weeks for each game; a score takes minutes. Run your tests only on the pitches with the best odds, and spend the same budget on fewer, stronger bets.

Our scouts already do this.

Great, this makes them faster. A first pass in a minute per pitch lets the same team cover ten times more pitches, and spend its own judgment on the strongest ones.

Which genres?

All Steam genres, in one model. We report accuracy separately for roguelikes, platformers and metroidvanias, simulation and management, idle and incremental, and co-op.

Send us the 10 pitches you are deciding on now

Ranked with forecasts within 24 hours. Free, under NDA.

Send 10 pitches