OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

Cosmo Edge

Cosmo Edge is a C++17 edge AI engine and console for deploying and operating video analytics, vision-language model and event workflows on Sophon, Rockchip and x86 hardware. It includes model import, visual orchestration, alarms and event delivery, with platform-specific builds, inference runtimes and model artifacts.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Cosmo Edge — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when video models need to become configurable edge applications, or when a team needs to build and operate visual detection and event workflows on supported hardware.

Who it is for

  • Teams deploying video analytics on supported edge hardware.
  • Developers assembling on-device vision-language and event-driven video workflows.
  • Engineers evaluating a video AI workflow locally on x86 before targeting an edge platform.

Documented uses

  • Deploy video AI models as edge applications.
  • Configure and operate video detection tasks and workflows.
  • Connect video analytics events with alarms and event delivery.
  • Evaluate a local x86 workflow before selecting a target NPU platform.
Before you choose

Requirements & limits

The repository is public and the core engine and console are released under Apache-2.0. The stable v1.1.0 release has packages; the documented local x86 quick start builds current main.

  • Select a target platform before deployment: Sophon, Rockchip and x86 use distinct runtimes, builds and model artifacts.
  • The documented v1.1.0 hardware scope includes BM1688, CV186X, RK3576 and RV1126B; platform-specific VLM claims differ, and RV1126B VLM is outside that release claim.
  • The local x86 route does not require edge hardware, but its throughput is lower than an NPU deployment.
  • Apple Silicon support is a separate amd64-emulation Preview for a local single-video workflow, not native performance evidence.
  • The Apache-2.0 scope covers the repository's core engine and console; certified hardware, commercial preset models and Model Guard distribution protection have separate boundaries.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start CosmoEdge locally on Linux x86 with Docker Compose

Build and start the CosmoEdge x86 services locally, then open the web interface.

Before you start

  • A Linux x86_64 host with Docker and Docker Compose installed.
  • Access to the CosmoEdge GitHub repository.

Steps

  1. Clone the repository with the command below and enter the cosmo-edge directory.

    Documented command
    git clone https://github.com/cosmo-wander-ai/cosmo-edge.git

  2. Start the services with the command below.

    Documented command
    docker compose -f docker-compose.x86.yml up -d --build

  3. Check service status with the command below.

    Documented command
    docker compose -f docker-compose.x86.yml ps

  4. Open http://127.0.0.1:8080 on the host.

Expected outcome

The x86 services are running and the local CosmoEdge interface is reachable at port 8080.

Guide limitations

  • The x86 mode has lower throughput than an NPU deployment.
  • The first build downloads dependencies and compiles the project; duration depends on network and host.
  • To access the interface remotely, use the host IP and allow TCP 8080 through the host firewall.
  • Starting the interface does not itself create or run a detection task; continue with the Scenario Configuration tutorial.
Documentation 1 ↗Documentation 2 ↗
Source-based practical guide

Launch the Apple Silicon macOS Docker Preview

Start the isolated amd64-emulation Docker Preview on an Apple Silicon Mac and open its local interface.

Before you start

  • An Apple Silicon Mac with the Docker Desktop Preview environment.
  • The CosmoEdge repository cloned locally.

Steps

  1. From the repository directory, run ./scripts/macos-docker-preview.sh doctor.

  2. Start the Preview with ./scripts/macos-docker-preview.sh up.

  3. Check its status with ./scripts/macos-docker-preview.sh status.

  4. Open http://127.0.0.1:8080 on the same Mac.

Expected outcome

The local Preview is started and its interface is reachable on the Mac at port 8080.

Guide limitations

  • This is a linux/amd64 emulation Preview for single-video local evaluation, not a native macOS binary or Sophon or Rockchip NPU deployment.
  • It is not production performance evidence.
  • Read the macOS Docker Preview scope before using this path.
Documentation 1 ↗Documentation 2 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Cosmo EdgeDeploy video AI models as edge applications.Linux, Self-hostedThe repository is public and the core engine and console are released under Apache-2.0. The stable v1.1.0 release has packages; the documented local x86 quick start builds current main.
MorphingdbImport a libtorch model and create predictions from values in a PostgreSQL table.Docker, Linux, Self-hosted, ServerThe README documents a Docker image and a build-from-repository option, followed by connecting to PostgreSQL and enabling the pgdl extension.
Gateway — pyadavSend chat-completion requests to a provider through a common server endpoint.ServerThe repository is public and identifies the project as Apache-2.0 licensed. README installation instructions include binaries, platform package managers and Docker.
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
cosmo-wander-ai/cosmo-edge · Official homepage title · retained in its source language
How we review sourcesSuggest a correction