GoogleGemini 4 ArgonVideo → Text

Text models are also served through an LLM proxy that speaks each vendor's own API. Requests and responses — streaming included — pass through exactly as the vendor defines them, so code written against the vendor's API or official SDK works here by changing only the base URL and the API key.

The same model is also reachable through one OpenAI-compatible Chat Completions endpoint that serves Claude, GPT and Gemini alike — switch vendors by changing model. When you'd rather use Picsart's own parameter names, use the unified API.

Authenticate with your Picsart API key, as Authorization: Bearer <key> (see Authentication). Any vendor credentials you send are ignored.

All on https://api.picsart.com, with gemini-4-argon as the model id.

One endpoint and one client for every text model, with the OpenAI SDK (pip install openai / npm install openai). The vendor is picked from model, so switching to another model means changing only that.

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["PICSART_API_KEY"],
    base_url="https://api.picsart.com/llm-proxy/openai/v1",
)

completion = client.chat.completions.create(
    model="gemini-4-argon",
    messages=[{"role": "user", "content": "Write a haiku about the ocean."}],
)

print(completion.choices[0].message.content)

The vendor's own generateContent API with the Google Gen AI SDK (pip install google-genai / npm install @google/genai) — the request bodies are the vendor's own, unchanged.

import os
from google import genai
from google.genai import types

client = genai.Client(
    api_key="unused",  # the SDK requires one; your Picsart key goes in the header below
    http_options=types.HttpOptions(
        base_url="https://api.picsart.com/llm-proxy/gemini",
        headers={"Authorization": f"Bearer {os.environ['PICSART_API_KEY']}"},
    ),
)

response = client.models.generate_content(
    model="gemini-4-argon",
    contents="Write a haiku about the ocean.",
)

print(response.text)