Create an AI influencer on one API. A month of posts for $5.53.

Example: one character, one month

Every image and clip bills, including the tries you discard. Story clips come from text prompts, so they show places and moods rather than your character.

Keep the same face

An image model has no memory of your character. The reference portrait is what carries the face from post to post.

• One reference portrait: Generate a clear, front-facing portrait and keep it. Send that same file with every post.
• Change the scene, not the person: Describe the place, outfit and light, and say "the same woman" or "the same man". Leave the face out of the prompt.
• Keep the ones that match: The model can drift. Budget two tries per post and publish the one that looks like your character.

Make a post from the reference

• Create the reference: Call /v1/images/generations with gpt-image-2 and a detailed description of your character. Save the portrait you like.
• Send it with each scene: POST the reference to /v1/images/edits with a prompt for the new scene and a 4:5 size.
• Write the caption: Ask a text model such as Claude Haiku 4.5 20251001 for a caption in your character's voice, on the same key.

import base64, requests

API = "https://api.apisrouter.com/v1"
headers = {"Authorization": "Bearer sk-your-apisrouter-key"}

post = requests.post(f"{API}/images/edits", headers=headers,
    files={"image": open("mia_reference.png", "rb")},
    data={
        "model": "gpt-image-2",
        "prompt": "The same woman, same face and hair, drinking coffee at a Lisbon cafe, candid, golden hour",
        "size": "2048x2560",
    },
).json()["data"][0]
data = requests.get(post["url"]).content if post.get("url") else base64.b64decode(post["b64_json"])
with open("post_01.png", "wb") as file:
    file.write(data)

Two ways to pay

Both fill one balance, spent per image, clip and token.

FAQ