gemini-web-quicker-skill

Access the Gemini Web API by routing queries through a Quicker action that runs a self-contained local HTTP service. This enables local agents to use Gemini Thinking, Flash/Pro, and Image models for multimodal processing.

Send image and video files to the Gemini Web API using the provided Python client, which supports base64-encoded media payloads and multiple endpoints for file uploads to handle multimodal inputs.

You can optimize video analysis with the Gemini API through the Python client's optional FFmpeg-based video compression feature, which automatically prepares and compresses video streams for processing.

Start the local HTTP service by launching the Quicker Gemini action to ensure the service runs on port 55557, then execute the python scripts/chat.py file with your prompt to begin querying.

Generate images locally using the Gemini 3 Image model via the Quicker action bridge, which supports streaming image generation responses directly through the self-hosted local HTTP API service.

The Gemini Web API bridge supports streaming responses for deep reasoning tasks, allowing local agents to process continuous outputs from Gemini 3.0 Thinking and Flash models for code and data analysis.