Vectorise.me has opened its image-to-SVG conversion engine to developers via a REST API and MCP server, enabling AI agents to vectorize images and export to multiple formats without local infrastructure. Pricing is per-conversion (starting at $0.0075 per image), with optional server-side AI restoration and preset profiles for different image types; the API is production-ready and compatible with Claude, Cursor, and other MCP clients.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Vectorise.me, an image-to-SVG conversion service, has released a REST API and an official MCP (Model Context Protocol) server that allows AI agents and applications to vectorize raster images and export them to multiple formats including PDF, EPS, DXF, PNG, JPG, WEBP, and AVIF. The service offers six scenario presets (auto, logo, illustration, photo, pixel_art, technical_drawing) and optional server-side edge recovery using real_esrgan for degraded sources or pixel_exact for pixel-art grids.
Why it matters
Developers building AI agents (compatible with Claude, Cursor, and other MCP clients) can now integrate image vectorization directly into workflows without managing local GPU infrastructure—the server handles all processing, including optional AI-powered restoration. Pricing is pay-per-conversion with credits (fast = 1 credit, balanced_crisp = 2, high_precision = 5), starting at $0.0075 per image, plus Pro plans with 500 included credits per month, making it accessible for both small scripts and high-volume pipelines.
What to watch
API rate limits are 200 requests per minute per key with per-key concurrent conversion caps. The service includes a self-describing OpenAPI 3.1 spec at /api/v1 for agent discovery, batch processing for up to 10 images per request (or 20 SVGs per export call on Pro), and error codes for queue/capacity/rate-limit scenarios; results are guaranteed to match the web app output since all three interfaces (web, REST, MCP) share the same underlying engine and settings.
Vectorise.me has announced a production-ready REST API and MCP server for converting raster images to SVG and exporting to multiple vector and raster formats. The REST API is accessed via simple X-API-Key headers (no OAuth or token refresh) and accepts JPG, PNG, BMP, WEBP, or AVIF files up to 15 MB. The core /api/v1/convert endpoint returns SVG markup and a color palette; the /api/v1/export endpoint converts SVG output to PDF, EPS, DXF, PNG, JPG, WEBP, AVIF, or back to SVG.
Six scenario presets tune the vectorization engine for different image types: auto (default, image-type detection), logo (sharp edges and exact brand colors), illustration (smooth curves), photo (tonal gradients), pixel_art (pixel-faithful grids), and technical_drawing (sharp black and white). Quality tiers (fast, balanced_crisp, high_precision) trade speed for precision and cost 1, 2, and 5 credits respectively. Optional server-side edge recovery—via edge_recovery=real_esrgan for 2× restoration of degraded sources or edge_recovery=pixel_exact for pixel-art grid recovery—runs on Vectorise.Me infrastructure, never client-side. The API includes batch endpoints (/api/v1/convert-batch for up to 10 images, /api/v1/export-batch for up to 20 SVGs on Pro plans) and diagnostics to show whether recovery actually executed (Diagnostics.edgeRecoveryResolved).
Pricing is per-conversion: credit packs are $5.99 for 500 credits, $14.99 for 1,500, and $29.99 for 4,000 (from $0.0075 per image). Pro subscribers receive 500 credits per month included. Rate limits are 200 requests per minute per API key, with X-RateLimit-* response headers and retry guidance via Retry-After. An MCP server (vectorizer-mcp, also published as vectorise-mcp) exposes tools (vectorize_image, vectorize_batch, export_svg, get_usage, get_capabilities, get_status) for Claude, Cursor, and other MCP clients; configuration requires only an VECTORISE_API_KEY environment variable in the client config. A self-describing OpenAPI 3.1 spec at /api/v1/openapi.json allows AI agents to discover and validate available parameters. All three interfaces—web app, REST API, and MCP—share the same underlying engine and settings, ensuring output consistency.
Vectorise.me addresses a practical gap in the AI agent ecosystem: image vectorization has traditionally required either browser-side WebGPU/GPU infrastructure or complex local setup, but many production pipelines and agents need a simple, stateless conversion endpoint. By exposing the same engine used in its web app through a REST API and MCP server, the service removes setup friction and ensures consistency—developers working with Claude, Cursor, or other MCP clients can invoke vectorization as a tool without managing dependencies or model downloads.
The pricing model—pay-per-conversion with no subscription required—appeals to both experimental use (credits are inexpensive enough for quick tests) and high-volume workloads (Pro tier for consistent throughput). The inclusion of scenario presets (logo, illustration, photo, pixel_art, technical_drawing, and auto-detect) reflects a quality-first philosophy; rather than expose low-level parameters, the API defaults to sensible profiles that match the web app's behavior, making it easier for agents to produce good results without extensive tuning. Optional server-side recovery (real_esrgan and pixel_exact) is explicit and off by default, preventing surprise GPU costs or latency spikes.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
