All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
0.1.0 - 2026-03-31
- Initial release of eai
- EAI-Min: Lightweight runtime for resource-constrained devices
- EAI-Framework: Industrial-grade AI platform with orchestrator and policy engine
- LLM Runtime: llama.cpp backend with 12 curated models (TinyLlama 80MB to Llama 3.2 8B)
- Agent Loop: Think-act-observe cycle with tool framework (MQTT, sensor, HTTP)
- Ebot Server: Interactive AI assistant at 192.168.1.100:8420
- 5 hardware tiers: Nano (≤256KB), Micro (≤16MB), Edge (≤1GB), Standard (≤8GB), Cloud
- Adaptive/Personalized AI: User preference store, feedback buffer, LoRA adapter support
- Runtime Training: load_adapter/unload_adapter, train_step, save_checkpoint
- Preference-Aware Agent: Injects top-10 preferences and feedback summary into prompts
- Built-in Tools: preference.set, preference.get, feedback.record, model.status
- Framework Adaptive Engine: Resource-gated training with RAM + CPU temp checks
- Profiles: smart-camera, industrial-gateway, robot-controller, mobile-edge, adaptive-edge
- EIPC Integration: Optional ENI→EAI communication bridge
- Complete CI/CD pipeline with nightly, weekly, EoSim sanity, and simulation test runs
- Full cross-platform support (Linux, Windows, macOS)
- ISO/IEC standards compliance documentation
- MIT license