Iris gives coding agents reliable screenshots

Iris targets a real weakness in coding-agent workflows: agents need visual context, but browser automation stacks are often too heavy for a single trustworthy screenshot.

• –The shared CLI and MCP engine keeps local developer workflows and agent workflows consistent
• –Inline image responses let agents inspect results immediately without managing temporary files
• –Smart waits and lazy-load scrolling improve capture reliability on modern, JavaScript-heavy sites
• –Selector-based framing is useful for focused UI verification, while device presets support responsive checks
• –Its narrow scope is a strength: Iris deliberately avoids interaction scripting, diffs, and review workflows