API Reference &
Automation Guide
Everything you need to automate profiles, connect AI models, and build workflows on top of OmertaBrowser. If you're unsure where to start — use the AI assistant in the bottom-right corner. It has full knowledge of this documentation and can write scripts for you on demand.
Installation
Windows — Installer
Download the OmertaBrowser-Setup.exe installer from the main page. Run it as administrator. OmertaBrowser will be installed to %LOCALAPPDATA%\OmertaBrowser.
Windows — PowerShell
macOS
Linux (Debian / Ubuntu)
Your First Profile
After installation, open OmertaBrowser. You'll be prompted to create your first profile. Each profile is a fully isolated browser identity. Give it a name, choose an OS fingerprint template, and assign a proxy.
JavaScript Automation API
OmertaBrowser exposes a omerta global object inside every automation script. Scripts run in a sandboxed V8 context with full access to the profile's browser session via Puppeteer-compatible APIs.
omerta.profile(id)
Acquires a handle to the specified profile. If the profile is not already running, it is launched automatically.
| Parameter | Type | Description |
|---|---|---|
| id | number | string | Profile ID or name. Use 'active' for the currently focused profile. |
session.humanize(options)
Enables human-behavior simulation for all subsequent interactions: randomized typing delays, organic mouse movement curves, scroll patterns, and random micro-pauses.
| Option | Type | Description |
|---|---|---|
| typing | [min, max] | Typing delay range in milliseconds. Default: [60, 180]. |
| scroll | 'organic' | 'linear' | Scroll behavior pattern. Default: 'organic'. |
| mouse | 'natural' | 'direct' | Mouse movement algorithm. Default: 'natural' (Bezier curves). |
| pauses | boolean | Inject random pauses to simulate reading. Default: true. |
| mistakes | boolean | Simulate occasional typing mistakes with corrections. Default: false. |
omerta.ai(config)
Creates an AI agent bound to an LLM of your choice. The agent can accept natural-language goals and execute them autonomously inside the given profile.
| Option | Type | Description |
|---|---|---|
| model | string | 'claude', 'gpt-4o', 'gemini', 'mistral', or 'custom'. |
| apiKey | string | Your model provider API key. Not required when using the built-in model. |
| endpoint | string | Optional custom OpenAI-compatible endpoint URL. |
| temperature | number | LLM temperature. Default: 0.4. |
| memory | boolean | Enable cross-session agent memory. Default: false. |
Scheduler
Scripts can be scheduled to run at specific times, intervals, or triggered by events such as proxy rotation, session expiry, or profile health changes.
Built-in AI Assistant
Every installation of OmertaBrowser (Capo tier and above) includes a built-in AI assistant. The assistant has full knowledge of the OmertaBrowser API, your profile configurations, and your installed scripts.
Ctrl + Shift + A from within the browser. Ask it to write a script, explain an error, or optimize your fingerprint setup.The assistant can:
— Write complete automation scripts from a plain-language description.
— Debug errors in existing scripts and suggest fixes.
— Explain API methods and parameters in plain language.
— Recommend fingerprint configurations for specific use cases.
— Generate proxy rotation strategies and warmup sequences.
Connecting External AI Models
In addition to the built-in assistant, you can connect your own API keys for GPT-4o, Claude 3.5, Gemini Ultra, Mistral, or any OpenAI-compatible endpoint. Go to Settings → AI Models → Add Model.
REST API — Authentication
The REST API allows external applications to control OmertaBrowser profiles, scripts, and sessions. Authentication uses Bearer tokens generated from Settings → API Keys.
GET /api/v1/profiles
Returns a paginated list of all profiles.
POST /api/v1/profiles/:id/run-script
Fingerprint Parameters
OmertaBrowser controls the following parameters on a per-profile basis. All values are isolated and persist across sessions unless manually randomized.
| Parameter | Type | Description |
|---|---|---|
| canvas | noise | block | real | Canvas 2D fingerprint noise injection strategy. |
| webgl_vendor | string | Spoofed WebGL renderer vendor string. |
| webgl_renderer | string | Spoofed GPU renderer name (e.g. "NVIDIA GTX 1080"). |
| audio_noise | float 0–1 | AudioContext oscillation noise level. |
| fonts | string[] | Enumerated font list visible to sites. |
| screen_width | number | Reported screen width in CSS pixels. |
| screen_height | number | Reported screen height in CSS pixels. |
| device_memory | number | navigator.deviceMemory value (GB). |
| hardware_concurrency | number | navigator.hardwareConcurrency (CPU thread count). |
| timezone | IANA string | Timezone identifier, e.g. Europe/Berlin. |
| locale | BCP 47 | Language and region code, e.g. de-DE. |
| webrtc | disable | spoof | real | WebRTC leak prevention mode. |
| user_agent | string | Full User-Agent string override. |
| platform | string | navigator.platform value. |
| do_not_track | 0 | 1 | null | DNT header value. |
Changelog
v2.1.0 — March 2026
Added built-in AI assistant with script generation. Autonomous agent mode added to omerta.ai().run(). GPU fingerprint spoofing expanded to 200+ GPU profiles. REST API endpoints stabilized.
v2.0.0 — January 2026
Major release. Full JavaScript API rewrite. External AI model support (GPT-4o, Claude, Gemini). Scheduler engine added. Parallel profile launch up to 100 simultaneous sessions.
v1.8.0 — October 2025
AudioContext noise injection. WebRTC triple-mode control. Cookie jar import/export. macOS ARM support.