Back to feed
Dev.to
Dev.to
6/27/2026
From "I Can't Click" to a Full Testing Harness: How We Built Playwright for the Terminal

From "I Can't Click" to a Full Testing Harness: How We Built Playwright for the Terminal

Short summary

The author built TTT, a Go-based terminal IDE, and solved a testing blocker: Claude Code can't interact with live TUI sessions. Starting with Lua plugin APIs, they evolved the solution to a --exec CLI flag supporting scripted interactions (click, key, type, screenshot, debug) with deterministic screen dimensions. This enables AI agents to drive editors programmatically for fast, deterministic testing without manual file creation.

  • •Problem: Claude Code couldn't interact with live terminal UI sessions for testing
  • •Solution: evolved from Lua plugins to a --exec CLI flag supporting scripted interactions
  • •Impact: enables AI agents to drive editors for fast, accurate testing without writing separate test files

Generated with AI, which can make mistakes.

#ai-tools#ai-agents#open-source
Read full article at Dev.to

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

Playwright can test Chrome extensions. So why does my AI agent still need my help?
Dev.to
The original title is "Getting Started with Claude Code and Playwright CLI: A Step-by-Step Guide"
Dev.to
How We Test an AI Product Without Burning Credit
Dev.to
The original title is "How I Made a Product Demo with Claude Code, Playwright, and FFmpeg"
Dev.to