[general]
name=OpenGeoAgent
qgisMinimumVersion=3.28
qgisMaximumVersion=4.99
description=GeoAgent-powered multimodal AI chatbot for QGIS projects.
version=1.2.0
author=Qiusheng Wu
email=giswqs@gmail.com

about=OpenGeoAgent provides a dockable multimodal chatbot interface for QGIS that builds on GeoAgent's QGIS tool surface.

    Features:
    - Dockable chat panel for asking questions about the current QGIS project
    - QGIS-safe GeoAgent tools for layer inspection, map navigation, data loading, selection, processing, and project operations
    - Confirmation-gated PyQGIS script execution fallback for QGIS API tasks without a dedicated tool
    - Provider and model controls for Bedrock, OpenAI, ChatGPT/Codex OAuth, Anthropic, Google Gemini, Ollama, and LiteLLM
    - Voice prompt recording with OpenAI speech-to-text transcription into the chat editor
    - Image-aware chat with clipboard paste support for providers and models that support vision inputs
    - Screenshot attachments from the map canvas, selected map regions, the QGIS window, and selected screen regions
    - Clickable image previews with save/export actions
    - Copy Markdown transcript and copy executed PyQGIS script actions
    - Sample prompts, prompt history with Up/Down, and Ctrl+Enter sending
    - Settings panel for API keys, hosts, model defaults, and dependency status
    - One-click dependency installer using uv with an isolated virtual environment
    - Lazy startup checks to keep opening the chat dock responsive
    - GitHub update checker
    - Local installation and packaging scripts

    Requirements:
    - GeoAgent and provider clients, installable from Settings > Dependencies
    - OPENAI_API_KEY for voice transcription, which uses the OpenAI transcription API and may incur API costs

tracker=https://github.com/opengeos/GeoAgent/issues
repository=https://github.com/opengeos/GeoAgent

hasProcessingProvider=no

tags=ai, agent, qgis, geospatial, chatbot, geoagent, llm, gemini, pyqgis, vision, multimodal, screenshot

homepage=https://github.com/opengeos/GeoAgent
category=Plugins
icon=icons/icon.svg

experimental=False
deprecated=False

changelog=
    1.2.0 - Added direct image generation support
        - Added OpenAI Images API tool integration for direct image generation
        - Added clickable generated-image thumbnails with full-resolution preview
        - Added image model selection and optional OpenAI org/project settings
        - Improved voice/API-key environment fallback handling

    1.1.0 - Improved voice input controls
        - Added configurable OpenAI transcription model selection
        - Added microphone button, voice shortcut, and recording/transcription indicators
        - Improved voice transcription status handling and prompt insertion behavior

    1.0.0 - Initial stable OpenGeoAgent release
        - Added clipboard image paste support for chat attachments
        - Added voice prompt recording with OpenAI speech-to-text transcription
        - Added screenshot capture for map canvas, selected map regions, QGIS window, and selected screen regions
        - Added image preview, save, and export actions
        - Added confirmation-gated PyQGIS script execution fallback
        - Added copy Markdown transcript and copy executed PyQGIS script actions
        - Improved chat dock startup with lazy imports and lightweight dependency checks

    0.1.0 - Initial OpenGeoAgent plugin
        - Added GeoAgent-backed QGIS chat dock
        - Added provider/model settings and credential fields
        - Added one-click dependency installer
        - Added update checker and packaging scripts
