Reading market data

Which developer API read answers which question about Roblox games - current players, many games at once, history, genres, the market, rising games.

Building on the platform / Building a market game

A game's servers read the same market data the panel shows, with the game's own key. Everything comes from what the platform's collector has already stored: no read asks Roblox, and a game the platform does not track is absent or 404. Each route has its own reference page; this page says which one answers which question.

Where the developer API's market data comes from

The collector reads Roblox for every tracked game on its tier's schedule and stores the readings in Postgres. The API answers a key's reads from its read cache, fresh for 60 seconds and replaced on the next read after the collector stores new market data. A Template game keeps one copy of each answer in MemoryStore for all its servers.

Market data from Roblox to a game's servers. The collector reads Roblox for each tracked game, every 5 minutes to once a week by its tier, and stores the readings in Postgres. The API answers a game key's reads from its read cache, which is fresh for 60 seconds and replaced after the collector stores new data. One server of a Template game fetches each answer and keeps it in the shared MemoryStore cache, where every other server of the game reads it. No read by a game key ever asks Roblox.

How many players a game has right now

Read the game's ccu from GET /developer/v1/games?ids=<id>, or from GET /developer/v1/games/:id when you also need its description, peak or placements. ccu is the players at the game's last reading, lastSnapshotAt.

The players of many games at once

Put up to 100 universe ids in one call: GET /developer/v1/games?ids=1001,1002,1003. Unknown ids are left out, the answer is sorted by players, and every id fits on one page. Never read one /games/:id per game. With the Template, keep to about 40 ids a call. Worked example: Building a market game.

A game's players over time

GET /developer/v1/games/:id/history?range=24h: buckets of 5 minutes for 6 or 24 hours, 30 minutes for 7 days, 2 hours for 30 days, a day for all time.

Players by genre

GET /developer/v1/genres: each first-level genre, its tracked games and their players now. Filter games by one with GET /developer/v1/games?genre=Simulation.

The whole market over time

GET /developer/v1/market?range=7d: players in the 1,000 most played games over the range, the 6 largest genres, and new games each of the last 12 weeks.

Rising and breakout games

GET /developer/v1/games?sort=breakout&limit=20 lists the games breaking out now; sort=growth_1h, growth_24h or velocity rank by recent growth, and maxAgeDays=30 keeps new games only.

The game the key belongs to

GET /developer/v1/key answers the game's id, its name and its limit per minute.

How often to read market data

At most once a minute per route and query, for the whole game, through one shared cache: the platform's own answers are cached for 60 seconds and change only when the collector stores new readings. The key's 120 requests a minute are shared by every server. The rules and the refusals: Errors, limits and polling.

What a game's key cannot read

Studio data (projects, contracts, points, shares, revenue, owned analytics), a game's feedback, GitHub activity, and any game the platform does not track. To have a game tracked, ask a studio owner to open its page in the panel.

Updated