Find similar imagesacross your dataset — no training, no labels.

import os, requests, numpy as np
from pathlib import Path

api_key = os.environ["MSF_API_KEY"]

def embed(path):
    r = requests.post(
        "https://api.msightflow.ai/v1/embed",
        headers={"Authorization": f"Bearer {api_key}"},
        files={"image": Path(path).read_bytes()},
    ).json()
    return np.array(r["vector"], dtype=np.float32)   # already L2-normalised

# Build a tiny index
vectors = {p.name: embed(p) for p in Path("dataset/").glob("*.jpg")}

# Query
query = embed("query.jpg")
scores = {name: float(query @ v) for name, v in vectors.items()}
for name, sim in sorted(scores.items(), key=lambda kv: -kv[1])[:5]:
    print(f"{sim:.3f}  {name}")

# Hosted search — no vector DB required for small datasets.
import os, requests

files = [("query", open("query.jpg", "rb"))]
for p in Path("dataset/").glob("*.jpg"):
    files.append(("dataset", open(p, "rb")))

resp = requests.post(
    "https://api.msightflow.ai/v1/embed/search",
    headers={"Authorization": f"Bearer {os.environ['MSF_API_KEY']}"},
    files=files,
    data={"top_k": "5"},
)
for hit in resp.json()["results"]:
    print(f"{hit['similarity']:.3f}  {hit['filename']}")

import fetch from "node-fetch";
import FormData from "form-data";
import fs from "fs";

const form = new FormData();
form.append("image", fs.createReadStream("query.jpg"));

const resp = await fetch("https://api.msightflow.ai/v1/embed", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.MSF_API_KEY}` },
  body: form,
});
const { vector } = await resp.json();
console.log("dim:", vector.length, "norm:", Math.sqrt(vector.reduce((s,v)=>s+v*v,0)));

curl -X POST https://api.msightflow.ai/v1/embed \
  -H "Authorization: Bearer $MSF_API_KEY" \
  -F "image=@query.jpg"