What problem does it solve? Calling Picsart's four REST APIs (Image, Video, GenAI, Variable Data) requires knowing dozens of exact tool names, enum values, and async polling patterns. This Skill maps user intent to the correct MCP tool, enforces verbatim parameter usage, and chains submit calls with their paired pollers so jobs actually complete. ## Core Features & Use Cases - Four API surfaces in one entry point: Image editing (background removal, upscale, effects), Video editing (trim, crop, concat, transcribe), GenAI generation (text-to-image, text-to-video, speech, stickers, logos), and Variable Data bulk rendering from Replay templates. - Async job handling: Every submit tool is paired with its correct -getresult poller, with explicit rules to never declare success before a terminal status. - Intent disambiguation: Decision tables route similar requests (e.g. enhance vs. prompt-based edit, image vs. video background removal) to the right API surface. - Use Case: A user asks to remove the background from a product photo, upscale it 4x, then generate a 5-second promo video from the result — the Skill reads the relevant reference docs, calls image-remove-background, image-ultra-upscale with polling, then genai-image2video with polling, and returns the final URLs. ## Quick Start Ask the agent to remove the background from an image URL using the Picsart API and return the result URL.