Timmy_Foundation/hermes-agent
15
0
Fork 0
Code Issues 287 Pull Requests 92 Actions 1 Packages Projects Releases Wiki Activity

feat: Wire Gemma 4 vision into browser_tool for screenshot analysis
All checks were successful
Lint / lint (pull_request) Successful in 8s
Details

Browse Source 
Default browser_vision screenshots to google/gemma-4-27b-it (Gemma 4
native multimodal) for reduced latency and unified text+vision model.

Resolution order for _get_vision_model():
1. BROWSER_VISION_MODEL env var (new, browser-specific override)
2. auxiliary.browser_vision.model in config.yaml (new config key)
3. AUXILIARY_VISION_MODEL env var (existing global vision override)
4. Default: google/gemma-4-27b-it

Backward compatibility: existing AUXILIARY_VISION_MODEL users are
unaffected — their override still flows through to browser_vision.

Also documents the new auxiliary.browser_vision config section in
cli-config.yaml.example and adds 14 unit tests covering the full
priority chain.

Fixes #816

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Alexander Whitestone
2026-04-21 17:12:58 -04:00
parent 12b5d9a7fd
commit 95bb842a21
3 changed files with 34 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tools/browser_tool.py
View File

@@ -806,7 +806,7 @@ BROWSER_TOOL_SCHEMAS = [
},
{
"name": "browser_vision",
"description": "Take a screenshot of the current page and analyze it with vision AI. Use this when you need to visually understand what's on the page - especially useful for CAPTCHAs, visual verification challenges, complex layouts, or when the text snapshot doesn't capture important visual information. Returns both the AI analysis and a screenshot_path that you can share with the user by including MEDIA:<screenshot_path> in your response. Requires browser_navigate to be called first.",
"description": "Take a screenshot of the current page and analyze it with vision AI (default: Gemma 4 multimodal). Use this when you need to visually understand what's on the page - especially useful for CAPTCHAs, visual verification challenges, complex layouts, or when the text snapshot doesn't capture important visual information. Returns both the AI analysis and a screenshot_path that you can share with the user by including MEDIA:<screenshot_path> in your response. Requires browser_navigate to be called first. Vision model can be overridden via BROWSER_VISION_MODEL env var or auxiliary.browser_vision.model in config.yaml.",
"parameters": {
"type": "object",
"properties": {