Agent Runs now available in the Vercel MCP and CLI
Vercel just shipped Agent Runs debugging for production AI workflows—trace reasoning, tool calls, and token usage without leaving the CLI.
Why it matters
Vercel is embedding agent observability directly into its deployment platform, letting builders debug and optimize AI agent behavior in production. This moves agent management from black-box to transparent, critical for teams shipping multi-agent systems at scale.
The key facts
6 to knowVercel MCP and CLI now support full agent trace inspection
Traces automatically ingested when deployed to Vercel
Four new CLI commands: list_agent_run_projects, list_agent_runs, get_agent_run, get_agent_run_trace
Traces include reasoning, tool calls, token usage, and subagent data
JSON output support for machine-readable traces
Enables agents to self-debug and update skills based on run data
Go to the source
Vercel Blogvercel.com
Publisher excerpt: Your agent can now inspect via the Vercel MCP and CLI for , the open-source agent framework.Agent Runseve eve traces are automatically ingested when deployed to Vercel and available as Agent Runs. The new and let you find projects with runs, list recent runs, and retrieve full traces, including…