Skip to Content

Turn your coding agent into a Vercel expert

One install loads a live knowledge graph and hand-picked skills into your agent of choice, so it’s always up to date with today’s docs.

Read the docs

Vercel Docsupdated 2m ago

Next.js Docsupdated 1h ago

AI SDK Docsupdated 2h ago

>Always currentYour agent reasons from current Vercel docs and conventions instead of its training cutoff. No more code against APIs that no longer exist.

Build an agentic data viz app Loading skills

4 of 28 skills loaded so far...

Catches mistakes earlyDeprecated packages and renamed APIs get flagged before they’re written, so the agent’s code passes review the first time.

Frequently asked questions

No. The plugin and the MCP server solve different problems and can run together in the same session. The plugin shapes how your agent thinks about Vercel by injecting platform knowledge and skills, while the Vercel MCP server gives your agent tools that perform actions on your account like listing projects, triggering deployments, and pulling logs.

The plugin installs into Claude Code, Cursor, Codex CLI, ChatGPT, Grok Build, VS Code, Kimi CLI, plus the GitHub Copilot CLI and app. One install command works across all of them, and support keeps expanding as more agents adopt the plugin format.

The plugin focuses on the workflows that produce the biggest accuracy gains for agents working on Vercel. Adding more skills increases token usage and raises the chance the agent loads guidance that doesn’t apply, so the current set targets Framework, AI, Infrastructure, Tooling, and Operations because those are where agents most often write code that needs correcting against current Vercel behavior.

The knowledge graph and the individual skills carry current guidance on renamed APIs, deprecated patterns, and sunset packages. When your agent loads a skill for the area it’s working in, that guidance steers it away from outdated references like @vercel/edge-config and toward the supported equivalent.

The SessionStart hook adds a one-time payload at the beginning of each session rather than running on every tool call, so the overhead lands in one place instead of accumulating per interaction. Token usage goes up when the agent loads a skill, but the tradeoff is more accurate Vercel code on the first attempt rather than output that needs manual correction afterward.

Yes. Route Claude Code, Cursor, and OpenAI Codex through AI Gateway by configuring the base URL and API key for each agent, and your spend across all three rolls into one dashboard with full request traces, provider fallback, and access to over 200 models. The AI Gateway integration is separate from the plugin, but the two work alongside each other and give teams both platform knowledge and unified billing in one setup.