#4949th of 80 in Agent platforms
cosmo-edge
C++ engine for video analytics and VLM tasks on Sophon, Rockchip and x86
- Stars
- 2.1k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- C
0 on 9 Jun 20262,092 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
CosmoEdge is a C++17 engine with a browser console for building video analytics pipelines: detection, tracking, zones, alarms, and VLM or GroundingDINO prompt-based checks. Each build targets one backend (Sophon BMRT, Rockchip RKNN, or ONNX Runtime on x86) and needs models converted for that chip. Events go out over REST, WebSocket, MQTT or HTTP webhooks, and an x86 Docker path runs the same UI without edge hardware.
Who it is for: Engineers deploying camera video analytics on Sophon or Rockchip edge boxes
Strengths
- Targets BM1688, CV186X, RK3576, RV1126B and x86 from one engine and console
- Built-in alarms, ONVIF and GB28181 camera access, MQTT and webhook event delivery
- x86 Docker compose path to try the full UI without edge hardware
- Published v1.1 benchmarks including a 72-hour run on four chips
Weaknesses
- Each build supports one backend; models must be converted per target chip
- CV186X and Rockchip v1.2 packages had no device validation in the release run
- Protected models, certified hardware and commercial preset models sit outside the Apache-2.0 code
- macOS preview runs under amd64 emulation, single-video trial only
What it needs
- GPU optional
- Docker + Compose
- Needs Docker, Docker Compose, Git
- Models: .nn (Sophon BMRT), .rknn (RKNN), RKLLM (RK3576), .onnx (ONNX Runtime), YOLO26
- port 8080
Quick answers
Is cosmo-edge open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Can you run cosmo-edge with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Does cosmo-edge need a GPU?
No. It runs on a CPU and can use a GPU when you have one.
Also in Agent platforms
See all 80| Rank | Project | Score |
|---|---|---|
| 1 | AutoGPTRole-based AI agents coordinated by one lead agent, plus a visual workflow builder | 95 out of 100 |
| 2 | daily_stock_analysisLLM-driven daily analysis of watchlist stocks across six markets, pushed to chat apps | 95 out of 100 |
| 3 | LangflowVisual builder for AI agents and workflows, exposed as API or MCP server | 94 out of 100 |
| #4 | deer-flowAgent harness running sub-agents, memory and sandboxes with pluggable skills | 93 out of 100 |
| #5 | DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs | 92 out of 100 |