Gemini API Documentation
by Google
The official reference for building on Gemini — function calling, structured output, long context, thinking, grounding and the Live API.
Overview
The Gemini API docs are the canonical reference for building against Google's Gemini model family, and they are organized by capability rather than by product. Core sections cover the quickstart and API key setup, the model catalog and how to choose between tiers, text/image/video generation, function calling for tool use, structured outputs for schema-constrained JSON, embeddings, long-context prompting, 'thinking' (extended reasoning) controls, the Files API for uploading documents and media, the Live API for low-latency bidirectional streaming, batch processing for high-volume offline jobs, grounding with Google Search for factual freshness, and agent-building patterns. Official SDKs are provided for Python (google.genai) and JavaScript (@google/genai), alongside plain REST. The current catalog spans Gemini 3.1 Pro for multimodal reasoning, Gemini 3.5 Flash as the frontier-class cost-effective workhorse, Gemini 3.1 Flash-Lite for high-volume cost-sensitive traffic, Nano Banana 2 and Pro for image generation and editing, Veo 3.1 for video with native audio, and Gemini Robotics for vision-language robotics. Access starts on a free tier via an API key, with paid tiers for higher rate limits and production use.
At a Glance
- Topic
- Models
- Level
- All Levels
- Format
- Documentation
- Cost
- Freemium
- Duration
- ~1 hour for the quickstart; reference docs are ongoing
- Provider
- Hands-on
- Yes — code/exercises
- Certificate
- None
What You’ll Learn
- ✓How to make your first Gemini API call in Python, JavaScript or REST and pick the right model tier for a workload
- ✓How to use function calling and structured outputs to get reliable, schema-conformant tool calls and JSON
- ✓How to exploit long context and 'thinking' controls, and what each costs you in latency and tokens
- ✓How to work with multimodal inputs and outputs — images, video, audio — via the Files API
- ✓How to ground responses in Google Search results for freshness and citations
- ✓How to build low-latency conversational experiences with the Live API and bulk jobs with batch processing
Highlights
- •First-party and continuously updated — the authoritative source on Gemini capabilities and limits
- •Covers the frontier-model features engineers actually ship: function calling, structured output, thinking, grounding, Live API, batch
- •Free tier available with an API key, so every example is runnable without a contract
- •Multimodal breadth in one API: text, image (Nano Banana), video (Veo), and robotics vision-language models
Who It’s For
Best For
- ✓AI engineers building applications or agents on Gemini
- ✓Teams comparing frontier model APIs on capabilities, context length and cost
- ✓Developers who need multimodal generation (image, video, audio) behind a single API
Prerequisites
- •Python or JavaScript, or comfort calling REST APIs
- •A Google AI Studio API key
- •Basic familiarity with prompting and LLM tool calling
FAQ
What is Gemini API Documentation?
Google's official developer documentation for the Gemini API, covering every capability of the Gemini model family from first API call to production agents. It is for AI engineers building on Gemini who need authoritative, current details on models, parameters and multimodal features rather than second-hand tutorials.
Is Gemini API Documentation free?
Gemini API Documentation offers free content, with paid options for certificates or premium features.
What level is Gemini API Documentation for?
Gemini API Documentation is aimed at a all levels audience. Recommended background: Python or JavaScript, or comfort calling REST APIs, A Google AI Studio API key, Basic familiarity with prompting and LLM tool calling.
How long does Gemini API Documentation take?
Expect roughly ~1 hour for the quickstart; reference docs are ongoing. Most learners work through it at their own pace.
What will I learn from Gemini API Documentation?
You'll learn: How to make your first Gemini API call in Python, JavaScript or REST and pick the right model tier for a workload; How to use function calling and structured outputs to get reliable, schema-conformant tool calls and JSON; How to exploit long context and 'thinking' controls, and what each costs you in latency and tokens; How to work with multimodal inputs and outputs — images, video, audio — via the Files API; How to ground responses in Google Search results for freshness and citations; How to build low-latency conversational experiences with the Live API and bulk jobs with batch processing.