Development logs made agent accessible.

How you use it

Wrap your dev command

The simplest path — add one line to your package.json:

One line, zero config. Every service gets its own log file plus a combined.log.

Add a framework plugin

For browser console capture, add the Vite or Next.js plugin:

They're complementary — use the CLI for server output and a plugin for browser console. Next.js + full setup →

try the playground →

How agents use it

agent-tail works best with AI tools that have access to your codebase — Claude Code, Cursor, Amp, and others. When your agent reads the log files, it gets:

• Timestamped errors with source locations
• Stack traces to trace the call path
• Server and browser output side by side
• The exact error message — no paraphrasing, no guessing

Without agent-tail

You copy-paste from browser devtools, describe the error in prose (“there's a 500 on the users page”), and hope the agent guesses right. Or you install an MCP browser tool that requires a running connection, can't be piped through grep, and gives you structured results you can't compose with other tools.

With agent-tail

Every time you start your dev server, agent-tail creates a new session directory and symlinks tmp/logs/latest/ to it. The agent runs grep ERROR tmp/logs/latest/ and gets the exact stack trace, source file, and line number. Plain files — no connection state, no tool registration, no token overhead. Agents already know how to read files.

Try it

Run a dummy command to see how simple the output capture really is. Or try the interactive playground →

npx agent-tail run 'server: echo "Hello world!"' && cat tmp/logs/latest/server.log

You run any command, output is captured to a log file, you or an agent can read the file. That's the full cycle.