Do not replace Nakama unless the thing you want to remove is operating Nakama.
Nakama is a capable open-source backend with realtime multiplayer, social features and an embedded server runtime. Crux makes a different trade: fewer extension points, no backend code runtime, and no Nakama cluster/database for your team to operate.
If you depend on Nakama runtime code, groups/chat, or its realtime multiplayer engine, staying on Nakama can be the correct decision.
Nakama OSS / Heroic Cloud
source + server runtime
realtime multiplayer
chat / groups / social
self-host or managed Nakama
Crux
managed fixed-scope API
persistent game data
progression + economy
matchmaking + registry
webhooks / live config
no embedded code runtime Compare the actual surfaces, not category labels.
Nakama's documentation and source are public. Crux exposes its OpenAPI contract and MIT client SDKs. You can inspect both before deciding which ownership model fits.
Live API status
A live check of the hosted API right now. No made-up uptime percentage and no hidden demo environment.
Check current statusPublic API contract
The hosted HTTP surface is documented as OpenAPI. Inspect the wire format before choosing an SDK or writing an adapter.
Open the specMIT client SDKs
The JavaScript, Godot, Unity, and Roblox client SDK source is public. The hosted backend itself remains a managed proprietary service.
Inspect the SDK sourceDocumented exit path
Read what is exportable today and the written commitment for advance notice plus an export window if the hosted service winds down.
Read the exit guaranteeMigration gap matrix
See which Nakama capabilities map to Crux, which change semantics, and which need to stay in another service before estimating a migration.
Inspect the gap matrixNakama wins when extensibility is the product requirement.
Heroic Labs documents Nakama as an open-source social/realtime backend with server runtime modules in Go, TypeScript/JavaScript and Lua. If that runtime is where your game rules live, replacing it with a narrower managed API can increase—not reduce—complexity.
Choose Nakama / Heroic Cloud when
- You need to run custom backend logic inside the platform.
- Realtime multiplayer, chat, groups/clans or richer social primitives are central.
- Self-hosting, private cloud or source-level control is mandatory.
- Your team already operates Nakama successfully, or wants Heroic Cloud to operate the same architecture.
Evaluate Crux when
- The backend is mostly standard game-service primitives rather than embedded runtime code.
- You want a managed API without running Nakama + database infrastructure.
- Authoritative game logic already lives in your game server or another service.
- You value a small explicit surface over maximum backend extensibility.
| Capability | Nakama | Crux |
|---|---|---|
| Source / self-host | Open source; self-host or Heroic Cloud | Hosted backend is proprietary; client SDKs are MIT and OpenAPI is public |
| Server runtime code | Go, TypeScript/JS and Lua runtime modules | No embedded runtime; logic runs in your game server/service |
| Realtime multiplayer | Built in | Not a netcode engine |
| Persistent player data | Storage engine | Player/project documents, stats and progression primitives |
| Social | Friends, groups, chat and more | Friends/requests/blocks; no groups or player chat service |
| Economy / leaderboards | Native APIs | Native APIs, but schemas/semantics are not 1:1 |
| Operations | You operate OSS, or Heroic Cloud operates Nakama | Managed Crux API; your game-server compute remains yours |
Prove the smaller surface is enough.
The right experiment is not “can Crux imitate every Nakama API?” It cannot. The useful experiment is whether your game actually needs the parts that do not map.
- Inventory Nakama runtime modules, realtime matches, chat/groups, storage, leaderboards, economy and authentication providers.
- Mark every dependency as native in Crux, semantic rewrite, or external service.
- Create the free Crux project and run one real request from your engine/server.
- Migrate one non-critical data path first; do not start with auth or the economy.
- Only plan a full cutover after the gap matrix has no unresolved critical feature.