The context switch
You're mid-build. Now you're alt-tabbing to find a stock photo.
You're writing a landing page, README, or blog post. You need a hero image. You leave your editor, open Unsplash, scroll 50 results, find nothing good, try Midjourney, spend 15 minutes on prompts, copy-paste the URL back, and lose your flow.
Your agent can read your code. It knows the context. It should be able to get you an image without you leaving.
Ask your agent. It searches OKSLOP, picks the right style, and inserts the image — inline, in your editor, in seconds. Same credits as okslop.com, no separate billing.
How it works
Your agent gets tools. You stay in flow.
Semantic search
"Cozy workspace photo" or "hero for a fintech landing page" — your agent searches by meaning, not keywords.
Generate on demand
When the library doesn't have it, generate it. Same prompt, same context — your agent handles the switch.
Context-aware picks
Your agent reads your code, understands the project, and picks images that fit — not random top results.
Insert inline
Image URLs drop directly into your code, README, or content. No copy-paste, no download-upload dance.
Setup
One install. Works everywhere.
Pick your agent
Claude Code, Cursor, Windsurf, Zed, v0 — we have an integration for each.
Install once
Claude: /plugin install. MCP clients: one line in config. v0: npx skill. Done.
Just ask
"Add a hero image to this landing page." Your agent handles search, selection, and insertion.
Ship
Commercial license included. No attribution required. The image is in your code, ready to deploy.
Why agents use this
Stock photo hunt vs. just ask.
| Approach | Time | Context loss | Result quality |
|---|---|---|---|
| Alt-tab to Unsplash | 10–20 min | High — you leave your editor | Generic — every project uses the same top results |
| Midjourney in Discord | 15–30 min prompting | Very high — different app entirely | AI-look unless you invest in prompt craft |
| Copy-paste URLs manually | 5 min per image | Medium — still switching tabs | Works but tedious at scale |
| Ask your agent (OKSLOP) | Under 10 seconds | Zero — stays in editor | Context-aware, distinct artist styles |
Integrations
Your agent, your choice. We support them all.
Claude Plugin
One install covers Claude Code, Cowork, and claude.ai. MCP server, skill, and slash commands bundled.
MCP Server
For Cursor, Windsurf, Zed, Continue, Cline — any tool that speaks MCP. One line in your config.
Agent Skill
For v0, Vercel agents, and agent-skills protocol tools. npx skill OKSLOP/okslop-skill.
CLI & SDK
npx okslop search/embed for scripts. TypeScript and Python SDKs for custom agent builds.
What agents build with this
Real workflows. Real time saved.
Landing pages
"Add a hero image that fits this SaaS landing page." Agent reads the copy, picks the mood, inserts the URL.
READMEs & docs
"Add a cover image for this open-source project." Consistent visuals across your documentation.
Blog posts
"Generate a header image for this article about developer productivity." Context-aware, style-matched.
Component libraries
Placeholder images, empty states, demo content — your agent populates them as it builds.
Agent FAQ
Questions developers ask.
Claude Plugin
One install for Claude Code, Cowork, and claude.ai. MCP server, skill, and slash commands bundled.
Install for ClaudeMCP Server
For Cursor, Windsurf, Zed, and other MCP clients. One line config.
Install MCP serverBuild your own
TypeScript and Python SDKs for custom agent integrations.
Read the docsStop alt-tabbing. Just ask.
Your agent searches 50K+ photos or generates custom images — inline, in your editor, in seconds. Same credits as okslop.com.