UAForge

Enterprise-grade, deterministic User Agent & Client Hint generator based on real-world browser statistics

Most "random user-agent" libraries are broken. They generate outdated browser strings, mix incompatible OS combinations, or lack the modern headers that fingerprinting systems now check.

UAForge takes a different approach. Instead of picking random strings, it simulates real users based on statistical probability. If Chrome 143 on Android holds 40% global market share, UAForge generates that identity 40% of the time.

It also generates matching Client Hints (Sec-CH-UA) headers automatically—allowing your automation to pass modern fingerprinting checks that go beyond the legacy User-Agent string.

Key Features

• Statistically Accurate — Weighted by real-world global usage data, updated weekly from caniuse.com
• Smart Correlations — Enforces valid browser↔OS mappings (no Safari on Windows)
• Real Hardware — Injects actual device models (Pixel 9, Galaxy S24, etc.) for mobile agents
• Client Hints — Generates Sec-CH-UA, Sec-CH-UA-Mobile, Sec-CH-UA-Platform, and GREASE tokens
• Deterministic — Seed support for consistent, reproducible identities across sessions

Installation

pip install git+https://github.com/sarperavci/uaforge.git

Note: The repo does not contain the market-share data files due to size constraints. They are automatically downloaded from the release assets while installing.

Quick Start

The API is designed to be simple. You generate an "Identity" object, which contains everything you need for your requests.

from uaforge.core.generator import UserAgentGenerator

# 1. Initialize the generator
agent = UserAgentGenerator()

# 2. Generate an identity
identity = agent.generate()

# 3. Get the headers (includes User-Agent AND Client Hints)
headers = identity.get_headers()
full_client_hints = identity.get_all_client_hints()

# Use with requests/httpx
# response = requests.get("https://httpbin.org/headers", headers=headers)

print(f"Browser: {identity.meta_browser.value}")
print(f"OS:      {identity.meta_os.value}")
print(headers)
print(full_client_hints)

Sample Output

Default Headers:

{
"User-Agent": "Mozilla/5.0 (Linux; Android 13; K) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/144.0.0.0 Mobile Safari/537.36",
"Sec-CH-UA": "\"Not A Brand\";v=\"99\", \"Chromium\";v=\"144\", \"Google Chrome\";v=\"144\"",
"Sec-CH-UA-Mobile": "?1",
"Sec-CH-UA-Platform": "\"Android\""
}

Full Client Hints:

{
"Sec-CH-UA": "\"Not A Brand\";v=\"99\", \"Chromium\";v=\"144\", \"Google Chrome\";v=\"144\"",
"Sec-CH-UA-Mobile": "?1",
"Sec-CH-UA-Platform": "\"Android\"",
"Sec-CH-UA-Full-Version-List": "\"Not A Brand\";v=\"99.0.0.0\", \"Chromium\";v=\"144.0.7559.59\", \"Google Chrome\";v=\"144.0.7559.59\"",
"Sec-CH-UA-Platform-Version": "13.0.0",
"Sec-CH-UA-Model": "\"Pixel 9\"",
"Sec-CH-UA-Arch": "\"arm\"",
"Sec-CH-UA-Bitness": "\"64\"",
"Sec-CH-UA-Full-Version": "\"144.0.7559.59\"",
"Sec-CH-UA-Form-Factors": "\"Mobile\"",
"Sec-CH-UA-WoW64": "?0",
"Sec-CH-Prefers-Color-Scheme": "dark"
}

Advanced Usage

Deterministic Generation (Sessions)

If you are managing long-running sessions, you need the User Agent to stay consistent across restarts. Use a seed.

# The identity generated here will ALWAYS be the same for seed 42
user = UserAgentGenerator(seed=42).generate()

print(user.user_agent) 
# Useful for associating a UA with a specific database UserID

Accessing Granular Data

Sometimes you need just the OS version or just the device model for analytics.

identity = agent.generate()

if identity.meta_device == "mobile":
    print(f"Device Model: {identity.ch_model}")  # e.g. "Pixel 8 Pro"
    print(f"Architecture: {identity.ch_arch}")   # e.g. "arm"

How it works

The Data Sources

We don't guess. We utilize three distinct data layers:

1. market_share.json: Global browser usage stats (Chrome, Safari, Edge, Firefox).
2. os_distribution.json: The probability of an OS given a specific browser (e.g., Safari is 100% macOS/iOS, but Chrome is split between Windows, Mac, Linux, and Android).
3. device_models.json: A curated list of ~500 real-world mobile device fingerprints.

Maintenance & Updates

The browser ecosystem moves fast. Market share data is automatically updated weekly via GitHub Actions by parsing the latest data from caniuse.com.

You can also trigger a manual update by running:

python scripts/parse_caniuse.py

Current Market Share Distribution

The table below shows the aggregated browser market share from the current dataset. Data is sourced from caniuse.com and updated automatically.

Last updated: 11-09-2026

License

MIT License. Feel free to use this in your commercial scrapers, bots or testing suites.