Back to directory

CosmoEdge

A C++ engine for deploying video analytics and on-device vision-language models to edge hardware such as Sophon, Rockchip and x86, with orchestration, alarms and benchmarking.

CosmoEdge

CosmoEdge is a C++ engine and application layer for turning video AI models into deployable edge applications across Sophon, Rockchip RKNN, and x86 hardware. Each supported platform uses its own runtime, build process and model artifacts, but shares a common orchestration, event, and observability workflow for video ingest, alarms and event delivery.

Version 1.1 adds a multi-platform release for BM1688, CV186X, RK3576 and RV1126B boards, a public benchmark pack covering single and concurrent workloads and a 72-hour dual-CV profile, an RKNN data path with DMA-BUF-to-RGA input and direct YOLOv8 tensor decoding, and a macOS Docker preview for local development on Apple Silicon.

The core engine and console are released under Apache-2.0; certified hardware, commercial preset models, and a model-encryption feature called Model Guard are separate, more restricted components.

More in Inference

See all 770
  • OpenDecision

    deepanwadhwa.github.io

    An open-source Python package that answers typed questions about application state or documents using a local natural-language-inference model, returning structured choice, score, yes/no or evidence-relation values instead of free text.

  • Lifeboat

    iterate.ai

    A local inference server that puts an OpenAI- and Anthropic-compatible API in front of open-weight language models running on hardware you control, from a laptop to a GPU cluster.

  • Claude S40

    github.com

    An unofficial Claude chat client for 2007-era Nokia Series 40 phones, built as a Java ME app plus a small Go relay server, with web search, reading mode and saved chats.