AI & ML APIs

Google Gemini API

Google's multimodal LLM API — text, vision, audio, and code generation with a genuinely free tier on Flash-class models and an OpenAI-compatible SDK for quick integration.

Use Cases

Build a free AI chat widget for a landing page using Gemini Flash and a serverless function

Prototype an image-captioning or OCR feature by sending user uploads to the vision model

Generate structured JSON or code snippets in an internal tool without paying for a heavy model

Free Tier

Free input & output tokens on Gemini 3.x Flash models, model-specific RPM/RPD limits, no credit card required

How to Maximize the Free Tier

The free tier gives you free input and output tokens on Flash-class models (Gemini 3.7/3.6/3.5 Flash) with rate limits set per model and per project — check your live numbers in Google AI Studio's rate-limit page since they change as models roll out. Daily quotas reset at midnight Pacific time, so schedule batch jobs after that boundary. The free tier has no spend-based cap, but your content is used to improve Google's products — keep that in mind before sending sensitive data, and never wire the free key into production. Start with Flash-Lite for high-volume tasks and reserve Flash for heavier reasoning. Combine with a serverless cache (Cloudflare Workers or Upstash) to dedupe repeated prompts and stretch your RPD. Upgrade to a billing account only when you need context caching, batch API (50% cheaper), or guaranteed throughput.

Getting Started

Go to aistudio.google.com → sign in with a Google account → create an API key (no credit card) → install the Google Gen AI SDK (`npm i @google/genai` or `pip install google-genai`) → call `client.models.generateContent` with a model like `gemini-3.7-flash` → check your usage and rate limits in the AI Studio dashboard.

Pros

Generous free quota: Free tokens on Flash-class models with no credit card — one of the most usable free LLM tiers available
Multimodal: Single API handles text, images, audio, and code — no need to stitch separate providers together
Strong ecosystem: AI Studio playground, SDKs for 10+ languages, and an OpenAI-compatible endpoint for easy migration

Cons

Data usage: Free tier content is used to improve Google products — not suitable for sensitive or proprietary data
Variable limits: RPM/RPD caps change per model and over time; you must check AI Studio to know your exact current quota
No production guarantee: Free tier offers no SLA and limited access to the most advanced models — production needs the paid tier

Alternatives

Hugging Face Inference API8/10★★★★☆Groq8/10★★★★☆Cohere7/10★★★½☆