PlanToCode Browser Bridge

Connects PlanToCode Desktop to requested Chrome tabs for page content, screenshots, and browser actions.

PlanToCode Browser Bridge connects Chrome to the local PlanToCode Desktop app through Chrome native messaging. It is for developers who want browser-task results, page text, Markdown, and screenshots returned to the PlanToCode run timeline. When you ask PlanToCode to use Chrome, the extension can open or reuse tabs, open Google Search results, extract requested web pages and return readable Markdown, extract page text including iframe content, inspect controls, click, type, scroll, select options, set checkbox and switch state, attach user-requested local files to file upload controls, take page screenshots, and close extension-created tabs. Results return to the PlanToCode run timeline so you can see what the agent saw: page text, Markdown, links, headings, screenshots, and browser action results. The bridge does not request cookie permissions and does not export cookies. Markdown extraction does not intentionally include current form input values, but screenshots and visible page content may include sensitive data shown on the page. When PlanToCode requests bridge status, the extension returns URLs and titles for open tabs so the task can target the intended tab. Chrome local storage holds connection status, metadata-only command, result, and cleanup summaries, and records for extension-created tabs and dedicated windows. Extracted page content, screenshots, and full browser command results are returned to PlanToCode Desktop and are not retained in Chrome local storage. PlanToCode Desktop must be installed for the extension to work. The desktop app registers the native messaging host com.plantocode.browser_bridge, receives browser commands from your PlanToCode run, and returns results to the run timeline. Main capabilities: - Open or reuse Chrome tabs for PlanToCode browser tasks. - Open Google Search results and return result links for follow-up inspection. - Extract requested web pages and return readable Markdown to PlanToCode. - Extract readable page text and Markdown from requested pages, including iframe content when Chrome permits frame access. - Inspect interactive elements before clicking or typing. - Click, type, press keys, select options, and scroll when commanded by the desktop app. - Set checkbox, radio, and switch controls to a requested checked state. - Attach validated local files to requested file upload controls without submitting the form. - Capture page screenshots for requested browser tasks. - Keep browser research in extension-owned tabs or dedicated windows when possible. - Close extension-owned tabs and conservatively remove exact history URLs that were created only by the extension.