# Plumerai

> Plumerai builds AI software for intelligent cameras, combining tiny on-device edge AI models with cloud Vision Language Models (VLMs). Products include person/vehicle/animal/package detection, Familiar Face Identification, cross-camera Re-Identification, noise-robust motion detection, VLM-based video search and custom AI notifications, plus the Plumerai Inference Engine — a deep-learning model compiler and runtime optimized for CPU inference on Arm Cortex-M / Cortex-A, RISC-V and x86_64. Plumerai exposes SDKs (C++, C, Python, Rust, Java) rather than a hosted REST API; the same unified API runs identically from microcontrollers to servers.

## Documentation
- [Documentation](https://docs.plumerai.com): Product and SDK documentation for Video Intelligence and the Inference Engine
- [C++ API reference](https://docs.plumerai.com/video_intelligence/cpp/): Video Intelligence C++ SDK
- [Python API reference](https://docs.plumerai.com/video_intelligence/python/): Video Intelligence Python SDK
- [Rust API reference](https://docs.plumerai.com/video_intelligence/rust/): Video Intelligence Rust SDK
- [Integration checklist](https://docs.plumerai.com/video_intelligence/integration_checklist/): Where new integrations should start

## Open-source libraries
- [larq](https://pypi.org/project/larq/): Training library for Binarized Neural Networks (pip install larq)
- [larq-compute-engine](https://pypi.org/project/larq-compute-engine/): Optimized BNN inference engine (pip install larq-compute-engine)
- [larq-zoo](https://pypi.org/project/larq-zoo/): Reference BNN implementations (pip install larq-zoo)
- [Larq GitHub](https://github.com/larq): Open-source BNN ecosystem maintained by Plumerai
- [larq.dev](https://larq.dev/): Larq ecosystem home

## Company
- [Website](https://plumerai.com/): Company home
- [Blog](https://blog.plumerai.com): Engineering and product blog
- [Contact](https://www.plumerai.com/contact_us): Contact / support
- [Careers](https://jobs.ashbyhq.com/plumerai): Open roles

## Notes
- No hosted REST/HTTP API, OpenAPI spec, or public /.well-known discovery surface is published; integration is via the on-device/server SDKs.