Develop with Gemini API

This skill extends AI assistants with direct access to Google's Gemini API, enabling advanced AI model capabilities including text generation, multimodal content processing, and structured data output. It allows assistants to leverage Gemini's chat, completion, summarization, image/audio/video/document understanding, function calling, JSON schema generation, Python code execution, context caching, and text embeddings for semantic search.

Use this skill when you need to generate conversational responses, summarize content, process multimodal inputs (images, audio, video, documents), invoke custom functions through the model, generate structured JSON matching specific schemas, execute Python code in a sandboxed environment, cache large contexts for efficiency, or create text embeddings for semantic search applications.

Do NOT use this skill if you require offline operation, as it depends on API connectivity to Google's services. Avoid using it when you need deterministic outputs for compliance or regulatory scenarios where model-generated variability is unacceptable.

Users provide text prompts, multimodal content (images, audio, video, documents), function definitions for the model to call, JSON schemas for structured output, Python code for execution, or text for embedding generation.

The skill returns generated text responses, chat completions, summaries, multimodal analysis results, function call requests with parameters, valid JSON matching provided schemas, Python code execution results, or text embeddings.

This skill serves developers building AI-powered applications that require advanced language model capabilities, data scientists needing multimodal content analysis, engineers implementing function-calling architectures where models trigger custom code, backend developers requiring structured JSON output for API integrations, and search engineers building semantic search systems with embeddings.