Models and Credits
Model inference is remote: EngineForge sends the prompt and selected working context through its gateway to the model you pick. Each request spends credits.
The model lineup
Models come from several third-party AI model providers. Some models require a paid plan; all paid plans include the full model lineup. The model picker shows the current lineup; in Plans & Models you can show or hide each model in the picker.
Request path
- You send a request from EngineForge.
- EngineForge assembles the prompt and the context needed for it.
- The request goes to the EngineForge gateway.
- The gateway routes it to the provider that serves the model you picked.
- The response streams back into the session.
Local versus remote
The project index and the engine connection stay on your computer. Index-derived or file context becomes remote data when it is included in a model request.
How credits work
Each plan includes monthly credits, shown as the percentage left this month. Every request spends some, and how much depends on the model: the most expensive model costs roughly a hundred times as much per token as the cheapest, so the same credits go much further on a lighter model.
Monthly credits reset each month and do not roll over. Usage top-ups are prepaid usage with no cash value; they never expire and are used only after the month’s credits run out. When both are empty, every model stops, including the cheapest.
What else uses credits
- Playtests, whether you send the Playtester from the card or Auto Playtest does.
- Generated sound effects, music and synthetic character voice lines.
- Generated game icons in Export Game. Choosing your own image is free.
Provider policies
Each model provider has its own terms for how long it keeps requests and whether it uses them. This page does not describe those terms; for privacy questions, contact support@engineforge.ai.