SingGuard

Name: SingGuard
Author: Ant Group

by Ant Group Vision · Moderation

SingGuard is Ant Group's open safety model for screening both text and images, built on Qwen vision-language models. It judges content against a risk policy, so one small model can check prompts, replies and generated images. The 0.8B size suits small GPUs and fast checks, while the 4B size catches more.

Publisher: Ant Group (opens in a new tab)
API model name: singguard:0.8b
Context window: 256K tokens
Sizes: 0.8B, 4B
Licence: Apache 2.0

Run SingGuard

Install it on a Pendra worker, then call it through our SDK with your pdr_sk_ API key.

Read the quickstart

from pendra import Pendra
import base64, pathlib

client = Pendra(api_key="pdr_sk_...")

image = base64.b64encode(pathlib.Path("photo.jpg").read_bytes()).decode()

result = client.moderations.create(
    model="singguard:0.8b",
    input=[
        {"type": "text", "text": "Profile photo for a new account"},
        {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{image}"}},
    ],
)

print(result.results[0].flagged)

import { Pendra } from "pendra";
import { readFileSync } from "node:fs";

const client = new Pendra({ apiKey: "pdr_sk_..." });

const image = readFileSync("photo.jpg").toString("base64");

const result = await client.moderations.create({
  model: "singguard:0.8b",
  input: [
    { type: "text", text: "Profile photo for a new account" },
    { type: "image_url", image_url: { url: `data:image/jpeg;base64,${image}` } },
  ],
});

console.log(result.results[0].flagged);

Run SingGuard on infrastructure you trust.

Deploy a worker, install SingGuard, and start serving it through one private API.

Try it free Read the quickstart