Use your Gemini subscription as an API

Ways to call Gemini from your own code

A Google AI Pro or Ultra plan doesn't give you an API key of its own. The table below compares the four ways to get Gemini into your own code.

On June 18, 2026, Gemini CLI and the Gemini Code Assist IDE extensions stopped serving Google AI Pro and Ultra, and Google moved those plans over to Antigravity CLI and the Antigravity app. If you reach for the Antigravity SDK for Python instead, its quickstart authenticates with a Gemini API key (GEMINI_API_KEY) or Google Cloud credentials for Gemini Enterprise, so it bills per token rather than using your plan.

Plan2API is the option that puts your Google AI plan behind a standard OpenAI-compatible endpoint.

Connect Google AI Pro or Ultra in 3 steps

1. Sign up at plan2api.com/connect with a Google account.
2. Connect Gemini by signing in with the Google account that has Google AI Pro or Ultra, on Google's own sign-in page.
3. Copy the p2a_ API key you get after connecting (the dashboard Quickstart page has copy-paste setup for curl, Python and Node).

Here is a streaming curl request to gemini/gemini-4-argon-high at https://api.plan2api.com/v1:

curl https://api.plan2api.com/v1/chat/completions \
  -H "Authorization: Bearer $PLAN2API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini/gemini-4-argon-high",
    "stream": true,
    "messages": [
      {"role": "user", "content": "Explain this stack trace in two sentences: ..."}
    ]
  }'

Gemini models you can call

Plan2API serves 5 Gemini models through Antigravity, each with a 1,048,576-token context window. The table below lists them next to what the same model costs per million tokens on the pay-per-token Gemini API.

• The gemini/ prefix picks Gemini, and google/ and antigravity/ work too.
• Every model takes text, image, audio and video input.
• GET /v1/models lists every model id Plan2API serves, including providers you haven't connected.

What works and what doesn't

• Requests use the OpenAI Chat Completions format on POST /v1/chat/completions.
• Streaming works.
• Tool and function calling work.
• Reasoning effort works via reasoning_effort or OpenRouter-style reasoning.effort.
• Image input works, and Plan2API fetches image URLs in messages for you (up to 32 per request, up to 20 MB and 15 seconds each).
• JSON mode works with response_format set to json_object or json_schema.
• temperature and top_p are passed through to Gemini.
• There is no /v1/responses, no /v1/embeddings and no native Gemini API format, only Chat Completions.
• The API sends Access-Control-Allow-Origin: * CORS headers, so browser apps can call it directly.

This example runs JSON mode on gemini/gemini-3.8-flash-high with the official OpenAI Python SDK:

from openai import OpenAI

client = OpenAI(
    base_url="https://api.plan2api.com/v1",
    api_key="p2a_...",
)

reply = client.chat.completions.create(
    model="gemini/gemini-3.8-flash-high",
    response_format={"type": "json_object"},
    messages=[
        {"role": "system", "content": "Reply with JSON: {\"sentiment\": ..., \"topics\": [...]}"},
        {"role": "user", "content": "The new release fixed sync, but exports are still slow."},
    ],
)
print(reply.choices[0].message.content)

Usage limits on Google AI Pro and Ultra

Plan2API adds no token quota of its own, so your Google AI plan's own limits apply. Plan2API reaches Gemini through Antigravity, where Google sets the limits in the table below.

Google says Antigravity limits track the amount of work the agent does, which means complex tasks use more quota, and that the limits can change to manage capacity. Google AI Plus and Free get a quota that refreshes weekly, while Plan2API connects Google AI Pro and Ultra.

If you have two or more Google AI accounts connected, a request that hits a limit moves to the next account before any output is sent. When every connected account is out of quota, the API returns HTTP 429 subscription_quota_exhausted with a Retry-After header set to the earliest reset, or 429 subscription_throttled when Google is rate limiting.

Use your Google AI plan in scripts and apps

Plan2API gives your Google AI plan a standard base URL (https://api.plan2api.com/v1) and one p2a_ key that can also route to Claude, ChatGPT (Codex) and Kimi subscriptions by model prefix, so the same OpenAI SDK code works across providers.

Connected logins are encrypted with HPKE (X25519) and AES-256-GCM so only your API key can unlock them, and the key's secret is never stored. Prompts and outputs are never stored, Plan2API's servers and database run on OVHcloud in Germany, and you sign in on Google's own page, so Plan2API never sees your Google password.

• Free is $0 with 1 connected subscription, 2 API keys, 50 chat requests a day and no card.
• Pro is $12 a month or $99 a year with 3 connected subscriptions, 10 API keys and no daily request cap.
• Scale is $29 a month or $249 a year with 25 connected subscriptions, 50 API keys and no daily request cap.