Empowers AI agents (Claude, Gemini, Antigravity) to search, filter, download, and organize millions of royalty-free images seamlessly.
imageHub-MCP is an authorial Model Context Protocol server developed by Alex Santos. It equips autonomous AI coding agents with 27 granular tools to discover, download, resize, and bookmark royalty-free visuals without leaving their conversational workflow.
How AI agents invoke imageHub-MCP tools through stdio / JSON-RPC to query and stream media assets.
Design choices, architectural patterns, and engineering decisions implemented for extreme reliability and developer experience.
Includes tools for keyword search, color filtering, orientation filters, category queries, multi-size asset downloads, favorites management, and markdown preview generation.
Every tool parameter is strictly validated using Zod, ensuring AI agents receive descriptive errors and self-correct their tool invocation payloads automatically.
Implements client-side query caching and verifies file existence before executing disk downloads, conserving bandwidth and keeping agent latency minimal.
Configurable via a simple JSON snippet in claude_desktop_config.json or antigravity MCP config, communicating over high-speed stdio streams.
Real-world usage, terminal configuration, and technical integration commands.
{
"mcpServers": {
"imagehub": {
"command": "npx",
"args": ["-y", "imagehub-mcp"],
"env": {
"PIXABAY_API_KEY": "YOUR_API_KEY"
}
}
}
}Key clarifications regarding licensing, compatibility, deployment, and security.
Yes. Pixabay provides free API keys with generous hourly allowances for personal and commercial usage.
Yes. Tools like download_image allow agents to save assets directly into targeted project directories.
Photos, vector illustrations, and transparent PNG illustrations.
Yes, licensed under the permissive MIT license.