Point your agent at Venice
One line, and there is nothing left to configure.
add and configure https://venice.ai/skill.md
That's the official Venice skill, versioned against the API spec. Nothing to install.
Supported today
Works in the tools you already run
Venice is OpenAI-compatible, so anything that takes a custom base URL takes Venice. Every card below opens its setup guide.
Your agent never needs another provider
All six run on one key and one base URL. See the full API for endpoints, pricing, and model discovery.
What your agent gains
Things your agent can't do alone
Web access
Skip the scraping stack
Your agent searches, scrapes pages into clean markdown, and cites what it used. Each is a flag on a request it already makes.
"venice_parameters": {
"enable_web_search": "on",
"enable_web_scraping": true,
"enable_web_citations": true
}
Your context isn't training data
Zero data retention. Your agent ships whole files and repo histories, and Venice keeps none of it.
No refusals mid-run
Venice hosts 63 open-weight text models with no policy layer on top. A refusal halfway through a loop strands your agent, and these don't send one.
New models the day they ship
Frontier launches go live on Venice the same day. Your agent reads the catalog at runtime instead of pinning an ID that rots.
Agent tooling
Three ways to connect
Pick whichever one your agent speaks.
Connect it now
Get a key, paste the line. That's the whole integration.