From a871948d8d4b0f774d4ec40467bab1078a9f28d5 Mon Sep 17 00:00:00 2001 From: kshitij <82637225+kshitijk4poor@users.noreply.github.com> Date: Wed, 12 Aug 2026 17:30:42 +0530 Subject: [PATCH] =?UTF-8?q?fix:=20correct=20Lightpanda=20fallback=20docs?= =?UTF-8?q?=20=E2=80=94=20remove=20nonexistent=20PDF/upload/clipboard=20ac?= =?UTF-8?q?tions?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Hermes has no browser PDF, file upload, or clipboard tools. The fallback mechanism only covers commands in _FALLBACK_ELIGIBLE (open, snapshot, screenshot, eval, click, fill, scroll, back, press, console, errors). The original docs described Lightpanda's general limitations, not Hermes's actual behavior. --- website/docs/user-guide/features/browser.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/website/docs/user-guide/features/browser.md b/website/docs/user-guide/features/browser.md index 7f2923d253162..4f7e36b16907a 100644 --- a/website/docs/user-guide/features/browser.md +++ b/website/docs/user-guide/features/browser.md @@ -353,7 +353,7 @@ AGENT_BROWSER_ENGINE=lightpanda Hermes drives Lightpanda through `agent-browser` over CDP, the same way it drives local Chrome. -**Automatic Chrome fallback.** Lightpanda doesn't yet cover everything Chrome does, so the integration is non-disruptive: Lightpanda handles the actions it supports, and Hermes transparently retries on Chrome for anything it doesn't. The supported set covers the core agent workflow — navigate, snapshot, click, type, scroll, back, press, and eval. Actions that fall back to Chrome include screenshots (Lightpanda has no graphical renderer), PDF generation, file uploads, and clipboard operations. Because Lightpanda skips visual rendering, `browser_vision` is pre-routed straight to Chrome. +**Automatic Chrome fallback.** Lightpanda doesn't yet cover everything Chrome does, so the integration is non-disruptive: Lightpanda handles the actions it supports, and Hermes transparently retries on Chrome for anything it doesn't. The supported set covers the core agent workflow — navigate, snapshot, click, type, scroll, back, press, and eval. Screenshots also fall back to Chrome because Lightpanda has no graphical renderer; `browser_vision` is pre-routed straight to Chrome for the same reason. ### Local Chromium-family browser via CDP (`/browser connect`)