Use the Gemini API without a Google Cloud project.
Gemini 2.0 Flash, Gemini 1.5 Pro, and Gemini 1.5 Flash β accessible from anywhere through JJAPI's OpenAI-compatible endpoint.
Latest models as of 2026-05-23
See all models βgemini-3.0-pro flagship Google's strongest current model with 2M-token context. The pick for long-document and multimodal workloads.
gemini-3.0-flash balanced Latest Flash. Cheapest per-token long-context model on the market.
gemini-2.5-pro flagship Late-2025 flagship with deep reasoning mode. Still excellent.
Why use JJAPI for this API
No GCP setup
Skip Google Cloud project creation, IAM, billing alerts, and the AI Studio waiting list.
Long context, low rate
Gemini's per-token rate at long context is the cheapest in the industry. We pass that through.
Native multimodal
Send images, audio, PDFs, video β Gemini accepts them all, and JJAPI relays them unchanged.
Free SSE streaming
Gemini's streaming format is translated to OpenAI SSE so your existing client works.
from openai import OpenAI
client = OpenAI(
base_url="https://api.jjapi.net/v1",
api_key="sk-jjapi-...",
)
r = client.chat.completions.create(
model="gemini-3.0-pro", # current as of 2026-05-23
messages=[{"role": "user", "content": "Hi"}],
)
print(r.choices[0].message.content)Frequently asked
Can I do 1M-token context windows?
Yes β gemini-1.5-pro on JJAPI supports the full 2M-token context. You pay per input token used, same as Google.
Does grounded search work?
Gemini's built-in google search grounding is exposed via the `tools` parameter. See the docs for the request shape.
What about Imagen for image generation?
Imagen-3 will be added soon. Today we route image generation to DALLΒ·E 3 / Flux by default.