Agents in Claude, ChatGPT and other cloud tools cannot open web links, so they used to pick photos by their file name. Now ViralHunt reads the pictures for them: the agent finds a licensed photo, looks at it, and sees the final post before handing it to you.
- Search licensed photos on Wikimedia Commons, with the author, the licence and a ready credit line taken from Commons itself.
- Look at any photo by its link before it goes into a template: does it show what the post says, is there a watermark, does it survive the crop.
- See the picture a draft will publish, exactly as it will go out.
- Upload a picture the agent made, straight from the conversation.
- See the template gallery as one picture with every template numbered, to choose one by its number.
Available in the Claude connector (viralhunt.io/mcp), the MCP server 1.1.36 and the skill 1.6.46. Nothing is published without your OK.