Merged
Conversation
- Introduced GroundedOpenAIChatAgent for separating visual grounding from reasoning. - Added GroundedComputerTool for resolving element descriptions to coordinates. - Implemented Grounder class for API calls to grounding models and coordinate parsing. - Created configuration and initialization for grounding models. - Updated Docker image version in browser agent example. - Added new grounding examples and configurations in the tools module.
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Contributor
|
📝 Documentation updates detected! New suggestion: Add comprehensive grounded agent architecture documentation for PR #113 |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Parth220
approved these changes
Sep 11, 2025
Co-authored-by: jaideep <jaideep@hud.so>
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Co-authored-by: jaideep <jaideep@hud.so>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces a new "grounded agent" architecture that separates visual grounding (element detection) from high-level reasoning, enabling more robust and modular browser automation. The main additions are a new agent class (
GroundedOpenAIChatAgent), supporting configuration and tooling for grounding, and an example script demonstrating usage. It also updates the browser agent Docker image version and adds support for OpenRouter API keys.Grounded Agent Architecture:
GroundedOpenAIChatAgentinhud/agents/grounded_openai.py, which uses a dedicated vision model for grounding UI elements and a planning model (e.g., GPT-4o) for reasoning. The agent exposes a synthetic "computer" tool, intercepts tool calls for grounding, and separates screenshot management.hud/tools/groundingmodule withGrounder,GroundedComputerTool, andGrounderConfigfor managing grounding model configuration and execution. [1] [2]Example and Configuration:
examples/grounded_agent.pyto demonstrate the grounded agent workflow, including environment setup, model configuration, and a sample form-filling task using separated vision and reasoning.hud/settings.pyto support theOPENROUTER_API_KEYfor OpenRouter-based vision models.Maintenance:
examples/03_browser_agent_loop.pyfrom0.1.2to0.1.3to ensure compatibility with the grounded agent.