Mcp
Wire your agent to DragonEye’s hosted APIs to run zero-shot video understanding—detection and classification—without training custom vision models first.
Overview
DragonEye MCP is a Build-phase MCP server that exposes zero-shot video object detection and category or attribute classification via https://nexus.api.dragoneye.ai/mcp.
What is this MCP server?
- Zero-shot video models for object detection without custom training datasets
- Category and attribute classification on video inputs via hosted Nexus API
- Streamable HTTP remote MCP at https://nexus.api.dragoneye.ai/mcp
- Suited to prototyping computer-vision features inside agent-driven build workflows
- Registry-listed server version 1.0.0 for catalog discovery
- MCP registry version 1.0.0
- 1 remote endpoint: https://nexus.api.dragoneye.ai/mcp
- Transport: streamable-http
What problem does it solve?
You need video understanding in your app but training and hosting custom vision models is too heavy for a solo builder timeline.
Who is it for?
Builders adding ML-lite video features to a SaaS or API who want agent-assisted integration to a hosted zero-shot vision backend.
Skip if: Projects that require fully offline inference, bespoke model fine-tuning only on private hardware, or non-video static-image-only stacks with no Nexus account path.
What do I get? / Deliverables
Your agent can invoke DragonEye tools over MCP to classify and detect objects in video during integration and feature prototyping.
- Registered remote MCP connection to DragonEye Nexus
- Agent-callable vision detection and classification tools in your workflow
- Faster prototype of video-understanding features in your product
Recommended MCP Servers
Journey fit
Build is where you integrate external AI capabilities into your product or internal tooling pipeline. Integrations is the right shelf because this MCP exposes a remote vision API for object detection and attribute classification, not local devops or shipping gates.
How it compares
Hosted computer-vision API over MCP, not a local agent skill or a generic FFmpeg utility.
Common Questions / FAQ
Who is DragonEye MCP for?
Developers and solo builders using MCP-enabled agents who need zero-shot video detection and classification without building their own vision training pipeline.
When should I use DragonEye MCP?
Use it during Build when you are integrating video upload, moderation, or analytics features and want the agent to call DragonEye classification and detection remotely.
How do I add DragonEye MCP to my agent?
Register the remote MCP URL https://nexus.api.dragoneye.ai/mcp as a streamable HTTP server in your client and authenticate per DragonEye’s Nexus API requirements.