Vercel StrawberryLast night’s failed build, explained before you sit down
A companion reads every overnight deployment, pulls the logs for the failures, quotes the actual error line, and drafts the ticket with the commit attached.
The deploy failed at 2:14 in the morning. By the time anyone looks there are three more deployments on top of it, the failure is four screens deep in a log, and the person who would recognise the error is in another timezone.
A companion has it done by nine. It lists the deployments across your projects, finds the failures, pulls each one’s build logs, and comes back with the error line rather than a status colour, plus which deployments since then succeeded.
| Dashboard and CLI | A Strawberry companion | |
|---|---|---|
| Reading build logs for a failed deploy | Yes, one deployment at a time | Across projects in one pass, quoting the error line |
| Correlating the failure with a commit or a changelog | You open the other tabs | Reads them in the session you are signed into |
| Turning last night’s failures into a ticket | You write it | Drafted with the log excerpt quoted in it |
| Checking whether a variable is missing on a project | Open project settings | Names listed in one call; values stay encrypted |
| Cancelling a running build | Yes | Yes, and it asks first. The only write in the surface |
See what happened to last night’s deploys
Ask what happened to last night’s deploys.
It finds the failures, pulls the logs, quotes the error line, and tells you whether you are debugging history or an outage.
The next step needs no new tool. The commit the failure points at is in the GitHub account you are signed into, the dependency in the stack trace has a changelog on the open web, and the incident belongs in your tracker. It reads all three and drafts the issue with the log excerpt quoted in it.
What can it change?
One thing.
Cancelling a build in progress, and it confirms first, because the build it stops might have been the fix.
Everything else reads: projects, deployments, build logs, domains, team details and environment variable names, with the values left encrypted. That is a connection you leave in place and stop thinking about.
- List projects and deployments; fetch a single deployment with its build logs.
- List domains registered on the account, and read team details.
- List environment variable names; values are not decrypted by this tool.
- Cancel an in-progress deployment, on request, after confirming.
Why does the answer sit outside Vercel?
Because "why did this break" usually ends in a commit, a package release, a status page, or a configuration change somebody made in an unrelated dashboard on Friday. Those are tabs, and a companion reads tabs.
It quotes the error line next to the commit and the dependency changelog, then writes the ticket from all three, which is the fifteen minutes at the start of every incident that nobody accounts for.
Get a deployment digest every morning
Run the same pass as a routine and it becomes a morning digest: overnight deployments by project, which failed, the error line for each, and whether the current production deployment is the one you expect. That is small operations work, finished before anybody sits down.
Deploying, promoting and rolling back stay with the person who has the authority to promote something. This pass runs before anybody decides what to ship.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents