Recraft API

Generates production-ready vector graphics (SVG-compatible) from natural language prompts using Recraft V4 model, enabling scalable graphics without quality loss at any resolution. The system interprets design intent from text descriptions and produces mathematically-defined vector paths suitable for logos, icons, and illustrations that can be infinitely scaled for print or digital use.

Generates high-fidelity raster images (PNG/JPEG) from text prompts with fine-grained style control, allowing specification of artistic direction, color palettes, and visual aesthetics. The API accepts style parameters and color specifications that constrain the generation process, producing photorealistic, illustrated, or stylized outputs matching brand guidelines or design specifications.

Exposes Recraft capabilities through the Model Context Protocol (MCP), enabling integration with MCP-compatible AI agents, IDEs, and applications. The MCP integration provides standardized tool definitions and schemas for image generation, editing, and processing operations, allowing AI systems to discover and invoke Recraft capabilities through a unified protocol without custom integration code.

Provides API key-based authentication for accessing Recraft API, with keys generated and managed through user profile dashboard. The authentication system issues unique API keys that authorize API requests, with keys retrievable from the user's profile section in the Recraft platform. This enables secure, per-user API access without sharing account credentials while maintaining audit trails of API usage.

Implements credit-based billing system where image generation consumes credits from monthly allocation, with credits resetting monthly and not rolling over to subsequent months. Users purchase subscription plans that include monthly credit allocations, with additional credits available through top-up purchases. This enables predictable monthly costs while preventing credit hoarding and encouraging regular usage.

Manages intellectual property rights and commercial usage permissions based on subscription tier, with free tier images owned by Recraft and publicly visible, while paid tier images owned by users with full commercial rights. The system tracks image ownership and usage rights, enabling users to determine whether generated images can be sold, republished, or used commercially based on their subscription level.

Enables users to earn free credits by referring other users to Recraft through shareable referral links. When referred users sign up and make purchases, the referrer receives credit rewards, creating a viral growth mechanism that incentivizes user acquisition. The system tracks referral relationships and automatically credits the referrer's account when referral conditions are met.

Generates images containing readable text of any length with exact positional control, allowing developers to specify where text elements appear within the composition. The API accepts text content and coordinate specifications, rendering typography that integrates naturally with visual elements rather than overlaying text post-generation, enabling creation of posters, social media graphics, and marketing materials with embedded messaging.

Provides access to multiple Recraft model versions (V4, V3, V2) allowing developers to select the appropriate model for their use case via API parameter. Each model version represents different capability levels and performance characteristics, enabling trade-offs between quality, speed, and cost. The API supports model selection at request time, allowing dynamic switching based on requirements without code changes.

Removes backgrounds from images (both AI-generated and user-uploaded) producing transparent PNG output suitable for compositing or product photography. The capability uses image analysis to identify foreground subjects and isolate them from backgrounds, enabling creation of clean product images, portrait cutouts, or design elements that can be placed on any background.

Enlarges images to higher resolutions while preserving or enhancing visual quality through AI-powered upscaling. The capability accepts raster images and target dimensions, applying machine learning-based super-resolution to reconstruct fine details and reduce artifacts, enabling use of lower-resolution generated images in high-resolution contexts (print, large displays) without quality degradation.

Converts raster images (PNG, JPEG) into vector graphics (SVG) through automated tracing and path reconstruction. The capability analyzes raster image data and generates mathematically-defined vector paths that approximate the original image, enabling scalable versions of raster content and integration with vector design workflows. This enables conversion of generated raster images into editable vector assets.

Removes or inpaints specific elements from images by specifying regions or objects to erase, using AI to intelligently fill removed areas with contextually appropriate content. The capability accepts images and region specifications (coordinates, masks, or object descriptions), analyzing surrounding context to generate plausible inpainted content that maintains visual coherence and aesthetic consistency.

Generates realistic product mockups by placing generated or uploaded images onto product templates (t-shirts, mugs, phone cases, etc.), creating photorealistic presentations of products in context. The capability accepts product images and template specifications, applying perspective transformation, lighting, and shadow effects to integrate the product image naturally into the mockup template, enabling e-commerce platforms and designers to visualize products without physical prototyping.

Enables specification and control of color palettes during image generation, allowing developers to constrain generated images to specific color schemes, brand colors, or aesthetic preferences. The API accepts color specifications (hex codes, RGB values, or palette names) and applies them as constraints during generation, ensuring output images adhere to color requirements without requiring post-processing color correction or iterative refinement.