Stocky

Overview

Stocky is a Model Context Protocol (MCP) server that bridges AI assistants with popular royalty‑free image repositories such as Pexels and Unsplash. By exposing a unified, async API, it eliminates the need for developers to write separate integration layers for each provider. The server accepts search queries, returns richly annotated results, and can download images in multiple sizes—all through a single MCP endpoint. This tight coupling with AI clients enables assistants to fetch, evaluate, and embed visual content directly into conversational flows or creative workflows.

The core problem Stocky solves is the fragmented nature of stock‑image services. Each provider exposes its own REST contract, rate limits, and data models, making it cumbersome for an AI assistant to offer seamless image discovery. Stocky abstracts these differences behind a consistent interface: , , and . Developers can therefore ask the assistant to “find a sunset beach photo” or “give me details for this image ID,” and the server handles provider selection, pagination, and error resilience automatically. The result is a frictionless experience where the assistant can present visual options without exposing API keys or dealing with provider quirks.

Key capabilities of Stocky include:

• Multi‑Provider Search – Simultaneously query Pexels and Unsplash, or target a single source when needed.
• Rich Metadata – Each result contains dimensions, photographer attribution, licensing info, and direct URLs for various image sizes.
• Pagination Support – Clients can request additional pages with simple and parameters, enabling large result sets to be browsed incrementally.
• Async Performance – All provider calls run concurrently, reducing latency for the assistant and keeping user interactions snappy.
• Graceful Error Handling – Network failures or rate‑limit hits are caught and returned as structured error messages, allowing the assistant to retry or fallback transparently.

Typical use cases span content creation, marketing automation, and educational tools. A graphic designer chatbot can search for imagery that matches a brand’s color palette, while an e‑commerce assistant might fetch product photos from multiple stock libraries to populate listings. In a data science workflow, an AI helper could pull visual examples for documentation or presentations without leaving the notebook environment.

Integration into MCP‑enabled pipelines is straightforward: once the Stocky server is registered in an AI client’s configuration, the assistant can invoke its tools via natural language prompts. The server’s JSON‑based responses map directly to the assistant’s knowledge graph, enabling chaining of image search with downstream tasks such as image editing or sentiment analysis. This tight coupling gives developers a powerful, reusable component that extends the visual intelligence of AI assistants across any platform that supports MCP.