API Documentation

Model Overview

Our domain valuation model is hosted on Replicate at humbleworth/price-predict-v1 . The model combines domain embeddings with word embeddings using Sentence Transformers to predict domain values across three sales channels.

Quick Start

The fastest way to get started is with Replicate's create-replicate tool.

npx create-replicate --model=humbleworth/price-predict-v1

Node.js

npm install replicate

export REPLICATE_API_TOKEN=r8_your_token_here

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

const output = await replicate.run(
  "humbleworth/price-predict-v1:a925db842c707850e4ca7b7e86b217692b0353a9ca05eb028802c4a85db93843",
  {
    input: {
      domains: "example.com, startup.net, innovation.org"
    }
  }
);

console.log(output);

Python

pip install replicate

import replicate

output = replicate.run(
    "humbleworth/price-predict-v1:a925db842c707850e4ca7b7e86b217692b0353a9ca05eb028802c4a85db93843",
    input={
        "domains": "example.com, startup.net, innovation.org"
    }
)

print(output)

Input Format

Pass domains as a comma-separated string in the domains input field. You can include up to 2,560 domains per request.

{
  "domains": "CoatFinder.com, CommissionWell.com, tech.net"
}

Response Format

The API returns a JSON object with an array of valuations. Each valuation includes estimated values across three sales channels.

{
  "valuations": [
    {
      "domain": "coatfinder.com",
      "auction": 35,
      "brokerage": 7871,
      "marketplace": 2616
    },
    {
      "domain": "commissionwell.com",
      "auction": 30,
      "brokerage": 9370,
      "marketplace": 3017
    }
  ]
}

Best Practices

• Submit domains in lowercase for consistency.
• Batch multiple domains (up to 2,560) for maximum efficiency—batching is more cost-effective than individual requests.
• Include TLD extensions (.com, .net, etc.) with each domain.
• Avoid special characters or internationalized domain names (IDNs).

Limitations & Considerations

• Estimates are AI-generated and should not replace human expertise for important decisions.
• Training data extends through early 2024; recent market trends may not be fully represented.
• The model is optimized for English-language domains.
• Trademark issues and legal considerations are not evaluated by the model.