Gleap AI agents can use Vercel through Composio, for teams that ship web apps on Vercel and hear from customers in Gleap. Kai Code and custom agents can read deployments and logs, and help a teammate confirm that a merged fix actually reached production before the customer hears back.
What a Gleap agent can do in Vercel
- Is the fix live: Get Deployments shows the latest production deployment and its commit, so the agent can tell a teammate whether a merged fix is in production.
- Build errors: Get Deployment Events returns the build output of a deployment, for example when a customer reports a page that stopped loading after a release.
- Runtime errors: Get Deployment Runtime Logs shows the function logs of a deployment around the time a Gleap bug report arrived.
- Checks before you answer: List Deployment Checks shows whether the checks on a deployment passed.
- Logs while fixing: with Vercel in Kai Code’s MCP settings, Kai Code can read build and runtime logs while it prepares a pull request for your engineers.
Example: confirming a fix before the reply
Your engineers merged Kai Code’s pull request for a customer’s upload bug yesterday, and Gleap asks a teammate to confirm the fix is live before the customer is told. The teammate asks the agent on the ticket to check Vercel. The agent finds the production deployment that includes the merge commit and sees that it finished without errors. The teammate confirms the release, and the customer gets the update on the original ticket.
How to connect Vercel
- In Gleap, open AI & Automation, then Agents, and select the custom agent that should use Vercel, or click Create agent.
- Under Tools, click Add tool, search for “Vercel” and select it.
- Click Connect Vercel and follow the authorization Gleap shows: a Vercel access token, or the client ID and secret of your own Vercel integration followed by a sign-in window.
- Open the Vercel tool and switch off every action the agent should not use, especially the environment variable reads, Create new deployment, Promote Deployment to Production and the delete actions.
- Publish and test the agent in chat. For Kai Code, add Vercel under Settings, MCP instead.
Kai Code is on Team, Pro and Enterprise, and Kai Support can use connected apps as tools on the same plans. Custom agents are on the Team, Pro and Enterprise plans.
FAQ
Does Gleap integrate with Vercel?
Yes. Custom agents and Kai Code use Vercel through Composio’s Vercel toolkit, which supports access tokens and OAuth. They read or change Vercel data only while they run; Gleap does not deploy anything on its own.
Does Gleap connect to Vercel through MCP?
Yes. Gleap connects to Composio’s MCP server for Vercel with the Vercel account you connect, and the agent calls only the actions you leave switched on. To use a different remote MCP server, add it by URL: as a Custom MCP tool for Kai Support or a custom agent, or as a custom HTTP server in Kai Code.
What can a Gleap AI agent do in Vercel?
It can list and read deployments, events, logs, checks, aliases, projects and domains, create deployments, promote them to production, manage aliases and domains, and read or change environment variables. Some environment variable actions return decrypted secrets, so switch them off for every agent.
Which Gleap plan includes the Vercel integration?
Kai Code can use Vercel on Team, Pro and Enterprise, and Kai Support can use connected apps as tools in its customer conversations on the same plans. Custom agents that use Vercel are part of the Team, Pro and Enterprise plans. The token’s scope decides which projects the agents see. AI usage is billed separately; see pricing.
Can the agent deploy or promote to production?
The toolkit can, which is why those actions should stay off. A merge is not a release: your pipeline ships the code, the agent can check what is live, and a teammate confirms the fix before the customer is told.
