Vercel, meet Reflectiv.

Keep an eye on your projects and deployments from a conversation.

What you can do

Connect Vercel to let your bots inspect projects, deployments, and build information. Supported deployment actions require an approval in Reflectiv before they run. Access is limited by the permissions and projects granted to the integration.

Connect your account

Open Reflectiv on your Mac and go to Marketplace → Vercel.
Add the integration, then select Connect with Vercel.
In your browser, sign in to Vercel and choose the account or team and projects you want to share.
Approve the requested permissions. Return to Reflectiv when the browser says authorization is complete.
Select Check connection to verify access.

Regular users do not need to create a Vercel integration or supply a client secret. The browser callback is https://auth.reflectiv.to/vercel/callback.

Try a conversation

“List the Vercel projects I’ve connected.”
“Check the latest deployment for our website.”
“Show me why that build failed.”

If you ask for a supported deployment change, review the project and action in the approval request before confirming.

Permissions and data

The integration requests project read access and deployment read/write access. Retrieved data, including requested logs, may be sent to your selected AI provider and saved in the conversation. Credentials are encrypted on your Mac. The hosted sign-in service temporarily handles the authorization exchange. See our Privacy Policy for details.

Disconnect or change access

Choose Disconnect in Reflectiv’s Vercel settings to remove the local connection. To revoke the integration’s access at Vercel, remove Reflectiv in your Vercel account or team’s integration settings. Reconnect to choose a different team or update your permissions.

Troubleshooting

Sign-in expired: start again from Reflectiv and complete authorization in the same browser.
No projects listed: check that you selected the correct team and granted access to the relevant projects.
Authorization unavailable: try again shortly or use the API-token option in the integration settings.
A deployment action failed: check its current status before retrying; a request may have completed even if the response was interrupted.

Need help? Contact us.