Desktop Pricing Docs Blog About GitHub View as Markdown (Agent) Get started
Human Agent
Upload. Transform. Deliver.

Image infrastructure
for agents.

Give your agents a way to upload, transform, and deliver images. Turn generated assets, screenshots, and product photos into CDN URLs through MCP, CLI, or API.

Working with a cloud agent that can't see your screen? Get the desktop app

Set up PixelVault so you can host images for me.
First tell me setup takes about a minute, then keep me posted as you go.
Open https://api.pixelvault.dev/connect?agent=<your name>&os=<mac|windows|linux|ios|android|web>&can=<what you can do: shell,browser,connectors,secrets> and follow what it says.
Can't open links? Run npx skills add pixelvault-dev/skill -g -y and use the pixelvault-upload skill.

Your agent has the image.
Now give it somewhere to go.

PixelVault handles image storage, transforms, and global delivery so your agent can keep building.

01 / CREATE

Generated asset → live website

Upload an image, create the sizes your page needs, and use the CDN URLs in your app.

02 / SHARE

Screenshot → shared context

Turn a local screenshot into a link a teammate or another agent can open.

03 / PREPARE

Product photo → ready to use

Remove a background, resize a photo, and deliver the variants from one image.

Your image. A real URL.

Upload an image and copy its hosted link. No account needed. Anonymous uploads are public and removed within 24 hours. If you’re signed in, uploads are saved to your library when possible; otherwise they use the temporary demo.

Next: transform an image →

Drag & drop an image here

or

or paste a screenshot (Ctrl/⌘+V)

PNG, JPEG, GIF, WebP, AVIF, SVG · up to 2 MB

Give your agent image tools.

Copy this prompt into your agent to get started, or connect your MCP client directly.

MCP connection instructions →

Read https://pixelvault.dev/llms.txt and help me set up PixelVault. Ask which image I want to upload, then upload it and return its CDN URL and a 400px-wide WebP variant.

  • Free to start
  • Try it without an account
Works with
Claude Code
Cursor
Windsurf
Copilot
Any HTTP client
Claude Code
Cursor
Windsurf
Copilot
Any HTTP client

Most image stacks are either low-level infra or enterprise tools that block autonomous agent workflows.

An agent-first image host
without the enterprise tax.

PixelVault gives agents and teams a clean contract: upload via API, get a CDN URL, and keep full control of domains and costs — a leaner alternative to Cloudinary and other image hosts.

⚡

API-first everything

Registration, API keys, uploads, usage — your agent can do it all with API calls. Full OpenAPI 3.1 spec, idempotent uploads, retry-safe by design.

🌐

Global CDN

Every image is served from a globally distributed CDN with enterprise-grade durability. Fast load times everywhere, automatic edge caching.

🤖

Agent-ready from day one

llms.txt, OpenAPI spec, .well-known/api-catalog. AI coding agents can discover, register, and start uploading without a human touching a browser.

🌎

$0 to start

200 MB storage, 500 uploads/month, 1 GB bandwidth. Enough for some businesses to run on—and deliberately so. No card or trial expiry. Upgrade when you need more capacity.

⚡

Zero egress fees

No bandwidth charges for serving images. Ever. Predictable pricing that doesn't punish growth.

🎨

On-the-fly transforms

Resize, crop, convert to WebP/AVIF, apply effects (blur, grayscale, rotate), and watermark — just add query params to any image URL, e.g. ?w=400&fit=cover&fmt=webp. Thumbnails, social cards, and edits in one request, no build pipeline. Every plan.

✂️

AI background removal

Turn any photo into a transparent cut-out with ?segment=foreground — subject isolated, background gone, no design tool. Add &background=white to composite it onto a solid fill. One URL, one request, on every plan.

🇪🇺

EU data residency

Your images and metadata are stored at rest in the EU — Cloudflare R2 (EU) plus an EU Postgres database. GDPR-friendly by default. Cookieless analytics, no tracking cookies, no consent banner.

📋

Paste & host, in your app

Add GitHub-style image handling to any textarea: users paste, drop, or pick an image and get a hosted CDN URL dropped in for them. One <script> tag (or a React/Vue hook) and a browser-safe publishable key — no upload backend to build. See the docs.

📦

No lock-in — export anytime

Pull every image in a project — plus a manifest.json — as a single tar archive, from the dashboard or pixelvault export. Free on every plan. Your data is always portable, so leaving is never a trap. See the docs.

🖥️

Desktop menu-bar app

Copy a screenshot — or hit a hotkey — and a hosted image URL lands on your clipboard, ready to paste into Codex cloud, Claude Code, or Cursor. For cloud agents that can't take a pasted image; nothing to install on the remote side. Free & open source. See the app.

Let your agent try it
before you commit.

AI agents can discover, register, and start uploading without a human touching a browser. Keys, uploads and usage are all API-first.

01
Discover
llms.txt + OpenAPI
02
Register
One POST, instant key
03
Upload
File → permanent CDN URL
04
Serve
Global edge, instant cache

Give your next image a home.

Connect your agent or build with the API. Free to start.

Read the quickstart