Intel® AI Gaming Assistant Library is an AI-powered local backend service that enables in-game intelligent assistance features, including screenshot understanding, multimodal retrieval, knowledge base Q&A, and memory management. It runs entirely on-device, leveraging Intel GPU and NPU for accelerated inference.
What's Inside?
- Vision — Screenshot-based scene recognition and image retrieval
- Knowledge Base — RAG pipeline with embedding, reranking, and LLM-based Q&A
- Memory — Persistent conversation memory management
- MMR — Multimodal retrieval combining vision and semantic search
The server exposes a local HTTP API (default port 9190) for client integration.
System Requirements
| Component | Requirement |
|---|---|
| OS | Windows 10/11 64-bit |
| CPU | Intel® Core™ Ultra processor (recommended) |
| GPU | Intel® Arc™ dGPU or Intel® Core™ Ultra integrated GPU |
| NPU | Intel® NPU (for vision models) |
| RAM | 16 GB minimum, 32 GB recommended |
| Disk | 20 GB free space |
Note: The LLM, embedding and rerank models are not bundled in this release.
Full Changelog: https://github.com/GameTechDev/IntelAIGamingAssistantLibrary/commits/v1.0.0