hub Vector Search Developer API

Semantic Search Bridge

Connect modern AI models, custom GPTs, and autonomous agents directly with the zeelproject.com catalog via high-speed semantic search.

manage_search API Specification (Text Search)

To run a text-based semantic search query, send an HTTP GET request to the search endpoint:

GET https://api.zeelproject.com/v1/?q={query}&page={page}&limit={limit}&lang={lang}&type={type}

Request Parameters

Parameter	Type	Status	Description
`q`	string	Required*	The search text query (e.g. `modern sofa`). *Optional only when performing image search via POST. Can be combined with image search for hybrid results.
`page`	integer	Optional	Page number for pagination. Default: `1`.
`limit`	integer	Optional	Quantity of results per page. Clamped from `5` to `40` (default: `5`).
`lang`	string	Optional	Output language translations. Supported: `en, ru, hy, fr, de, it, es, zh` (default: `en`).
`score`	float	Optional	Similarity score threshold filtering (from `0.0` to `1.0`, default: `0.0` / disabled).
`type`	string	Multi	Content type filter: `product` (3D Models + Textures), `model`, `scene`, `texture`, `set`. Supports comma-separated multi-filtering (e.g. `model,texture`). Omit to search all types.

camera_enhance Visual Search (Image, URL, or Base64)

Search the catalog visually using reference images. Built for AI models, agents, and modern apps with multiple flexible input methods:

GET https://api.zeelproject.com/v1/?image_url={url}&page={page}&limit={limit}&lang={lang}&type={type}

How Image Search Works

Send Image

Provide a public image_url, an uploaded image file, or image_base64.

CLIP Encoding

Image is processed through CLIP visual encoder to compute a high-dimensional embedding vector.

Qdrant Matching

Vector is matched in the vector database using Cosine similarity metric in milliseconds.

Data Enrichment

Results are enriched from SQL database with verified titles, prices, image URLs, and model links.

Visual Search Parameters

Parameter	Type / Method	Status	Description
`image_url`	string (URL) • GET / POST	Supported	Public URL of reference image (Recommended for ChatGPT & Gemini Actions).
`image` (or `file`)	file (binary) • POST multipart	Supported	Direct binary file upload (supports JPEG, PNG, WebP).
`image_base64`	string (Base64) • POST JSON	Supported	Base64-encoded image string in JSON payload.
`q`	string • GET / POST	Hybrid	Optional text query to combine with image for hybrid image+text search. Vectors are averaged and normalized.

terminal Interactive Sandbox

Run test queries live and inspect the structured JSON response payload below.

Search Mode

Limit

Query Parameter (q)

Page

Language (lang)

Score Threshold

Type (type)

// Responses will be displayed here in pretty-printed JSON...

Gateway: api.zeelproject.com Format: JSON (UTF-8)

integration_instructions Integration Code Snippets

Select your language to view ready-to-use code snippets for query integration:

1. Text Search Integration (HTTP GET)

curl -G "https://api.zeelproject.com/v1/" \
  --data-urlencode "q=modern sofa" \
  -d "page=1" \
  -d "limit=5" \
  -d "lang=en" \
  -d "score=0.295" \
  -d "type=product"

const url = new URL("https://api.zeelproject.com/v1/");
url.searchParams.append("q", "modern sofa");
url.searchParams.append("page", "1");
url.searchParams.append("limit", "5");
url.searchParams.append("lang", "en");
url.searchParams.append("score", "0.295");
url.searchParams.append("type", "product");

fetch(url)
  .then(res => res.json())
  .then(data => console.log(data))
  .catch(err => console.error(err));

import requests

url = "https://api.zeelproject.com/v1/"
params = {
    "q": "modern sofa",
    "page": 1,
    "limit": 5,
    "lang": "en",
    "score": 0.295,
    "type": "product"
}

response = requests.get(url, params=params)
print(response.json())

<?php
$url = "https://api.zeelproject.com/v1/?" . http_build_query([
    "q" => "modern sofa",
    "page" => 1,
    "limit" => 5,
    "lang" => "en",
    "score" => 0.295,
    "type" => "product"
]);

$ch = curl_init($url);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$response = curl_exec($ch);
$data = json_decode($response, true);
curl_close($ch);

print_r($data);
?>

2. Visual & Hybrid Search Integration (URL, File Upload, Base64 & Combo)

# Option 1: Image URL (GET - Supports optional 'q' for hybrid search)
curl -G "https://api.zeelproject.com/v1/" \
  --data-urlencode "image_url=https://example.com/photo.jpg" \
  --data-urlencode "q=red modern armchair" \
  -d "limit=5" \
  -d "type=model,texture"

# Option 2: Direct File Upload (POST multipart/form-data + optional 'q')
curl -X POST "https://api.zeelproject.com/v1/?q=red+modern+armchair&limit=5&type=product" \
  -F "image=@/path/to/photo.jpg"

# Option 3: Base64 JSON Payload (POST application/json + optional 'q')
curl -X POST "https://api.zeelproject.com/v1/" \
  -H "Content-Type: application/json" \
  -d '{"image_base64": "data:image/jpeg;base64,...", "q": "red modern armchair", "limit": 5, "type": "product"}'

// Option 1: Search by Image URL (GET + Optional Text)
const url = new URL("https://api.zeelproject.com/v1/");
url.searchParams.append("image_url", "https://example.com/photo.jpg");
url.searchParams.append("q", "red modern armchair"); // Optional hybrid text prompt
url.searchParams.append("limit", "5");
url.searchParams.append("type", "model,texture");
const resUrl = await fetch(url).then(r => r.json());

// Option 2: Direct File Upload (POST FormData + Optional Text)
const formData = new FormData();
formData.append("image", fileInput.files[0]);
const resFile = await fetch("https://api.zeelproject.com/v1/?q=red+modern+armchair&limit=5&type=product", {
  method: "POST",
  body: formData
}).then(r => r.json());

// Option 3: Base64 JSON Payload (POST JSON + Optional Text)
const resB64 = await fetch("https://api.zeelproject.com/v1/", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    image_base64: "data:image/jpeg;base64,...",
    q: "red modern armchair", // Optional hybrid text prompt
    limit: 5,
    type: "product"
  })
}).then(r => r.json());

import requests

# Option 1: Search by Image URL (GET + Optional Text)
res_url = requests.get("https://api.zeelproject.com/v1/", params={
    "image_url": "https://example.com/photo.jpg",
    "q": "red modern armchair", # Optional hybrid text prompt
    "limit": 5,
    "type": "model,texture"
})
print(res_url.json())

# Option 2: Direct File Upload (POST multipart + Optional Text)
with open("photo.jpg", "rb") as f:
    res_file = requests.post(
        "https://api.zeelproject.com/v1/",
        params={"q": "red modern armchair", "limit": 5, "type": "product"},
        files={"image": f}
    )
print(res_file.json())

# Option 3: Base64 JSON (POST application/json + Optional Text)
res_b64 = requests.post("https://api.zeelproject.com/v1/", json={
    "image_base64": "data:image/jpeg;base64,...",
    "q": "red modern armchair", # Optional hybrid text prompt
    "limit": 5,
    "type": "product"
})
print(res_b64.json())

<?php
// Option 1: Search by Image URL (GET + Optional Text)
$url = "https://api.zeelproject.com/v1/?" . http_build_query([
    "image_url" => "https://example.com/photo.jpg",
    "q"         => "red modern armchair", // Optional hybrid text prompt
    "limit"     => 5,
    "type"      => "model,texture"
]);
$data_url = json_decode(file_get_contents($url), true);

// Option 2: Direct File Upload (POST with CURLFile + Optional Text)
$ch = curl_init("https://api.zeelproject.com/v1/?q=" . urlencode("red modern armchair") . "&limit=5&type=product");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_POST, true);
curl_setopt($ch, CURLOPT_POSTFIELDS, [
    "image" => new CURLFile("/path/to/photo.jpg", "image/jpeg", "photo.jpg")
]);
$data_file = json_decode(curl_exec($ch), true);
curl_close($ch);

// Option 3: Base64 JSON Payload (POST JSON + Optional Text)
$ch = curl_init("https://api.zeelproject.com/v1/");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_POST, true);
curl_setopt($ch, CURLOPT_HTTPHEADER, ["Content-Type: application/json"]);
curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode([
    "image_base64" => "data:image/jpeg;base64,...",
    "q"            => "red modern armchair", // Optional hybrid text prompt
    "limit"        => 5,
    "type"         => "product"
]));
$data_b64 = json_decode(curl_exec($ch), true);
curl_close($ch);
?>