MCP SERVEREARLY ACCESS
Deepfake detection for AI agents.
Model Context Protocol server that gives any AI agent the ability to verify media authenticity. Built for Claude, GPT, and any MCP-compatible client.
HOW IT WORKS
How it works.
01
Add Witness to your MCP config
One JSON entry in your MCP configuration file. No separate installation.
02
Your agent calls detect_image or detect_video
The MCP server exposes clean tool definitions. Your agent decides when and what to verify.
03
Receive structured results
Verdict, confidence, and detailed signals returned as structured data the agent can reason about.
04
Build verification into any workflow
Research agents, content moderation pipelines, newsroom tools — any agent that processes media can verify it.
FAQ