Lahyer AI
Explanations for failed builds and logs, and an assistant that knows your projects and asks before it acts.
Lahyer AI works from what Lahyer knows about your projects: deployments, build and runtime logs, usage, domains and settings. It reads; it does not change anything on its own.
Explain a failed build
On a failed deployment, choose Explain with Lahyer AI. You get a short summary, the probable cause with the log lines behind it, and a suggested fix.
Explain logs
On the project's Logs tab, choose Explain with Lahyer AI to have it read the lines in view. Add a question first to steer it, such as "why are requests to /api/orders slow?".
Ask about a project, or all of them
Choose Ask AI on a project, or open its AI tab, to ask about that project: why the last deployment failed, what changed this week, why it feels slow. For speed questions it runs a performance review covering response times over the last week, cold starts, caching headers, image sizes, build time and database load, and lists what to fix first.
The AI page in the dashboard answers questions about every project in the organization: which one uses the most, what failed this week, how close you are to your plan's limits.
From the terminal, lahyer ask does the same. See the CLI.
It asks before it acts
Lahyer AI can propose a redeploy, a rollback, a promotion, a settings change or an environment variable change. Nothing happens until a person chooses Approve in the dashboard (or runs lahyer ask --approve); Dismiss drops it. Values for an environment variable change are typed in when you approve, not given to Lahyer AI.
Weekly digest
Every Monday, owners and admins get an email summing up the organization's week: deployments, traffic, build minutes and plan usage, with anything worth a look.
Daily allowance
Each member can make 25 Lahyer AI requests a day on Free and 200 on Pro, counted per organization. The allowance resets at 00:00 UTC.
Lahyer Studio is metered separately, in credits. See Lahyer Studio.