Pricing Model
Enigma charges for two components:
- Token Usage — AI model input/output tokens
- Compute Time — Browser instance runtime
Both are tracked per-task and returned in the usage object of each response.
Token Pricing
Enigma uses proprietary browser-optimized AI models:
| Model | Input (per 1M tokens) | Output (per 1M tokens) | Best For |
|---|
| enigma-browser-pro | $0.60 | $2.70 | Complex tasks, high accuracy |
| enigma-browser-fast | $$0.60 | $1.80 | Simple tasks, faster execution |
What Counts as Tokens?
Input Tokens:
- Your task description
- Page content (HTML, text visible to agent)
- Agent’s internal reasoning
- Previous task context in multi-task sessions
Output Tokens:
- Agent’s thoughts and actions
- Result message returned to you
- Navigation decisions
- Error messages
Compute Pricing
Compute time is charged per second of active session time:
| Resource | Rate |
|---|
| Browser Instance | $0.XX per minute |
- Time from session creation to termination
- Includes idle time between tasks
- Maximum 5 minutes per session
Avoid Idle Charges: Sessions remain billable until terminated. Always use terminateOnCompletion: true on your final task or manually terminate sessions when done.
Cost Examples
Typical costs for common tasks:
| Task Type | Tokens Used | Duration | Est. Cost |
|---|
| Simple search | ~10,000 | 20s | $0.01 |
| Form filling | ~25,000 | 45s | $0.03 |
| Data extraction | ~40,000 | 60s | $0.05 |
| Multi-step workflow | ~100,000 | 3min | $0.15 |
| Complex automation | ~200,000 | 5min | $0.30 |
These are estimates. Actual costs depend on page complexity, task specificity, and agent efficiency.
Cost Breakdown in Responses
Every task response includes detailed cost information:
{
"type": "task_completed",
"data": {
"message": "Task completed successfully",
"prompt_tokens": 12450,
"completion_tokens": 3200,
"total_tokens": 15650,
"completion_time": 23.5
},
"usage": {
"inputTokens": 12450,
"outputTokens": 3200,
"computeTime": 5,
"cost": 0.0124
}
}
| Field | Description |
|---|
prompt_tokens | Input tokens used |
completion_tokens | Output tokens generated |
total_tokens | Sum of input + output |
completion_time | Task duration in seconds |
computeTime | Billable compute time in seconds |
cost | Total cost in USD |
Cost Optimization Tips
1. Use terminateOnCompletion: true
Problem: Sessions continue running (and billing) after tasks complete.
Solution: Auto-terminate when done.
{
"taskDetails": "Search for products",
"terminateOnCompletion": true
}
2. Set Appropriate maxDuration
Problem: Default 5-minute sessions may run longer than needed.
Solution: Set realistic timeouts.
{
"taskDetails": "Quick search task",
"maxDuration": 60000 // 1 minute
}
3. Use enigma-browser-fast for Simple Tasks
Problem: Using the standard model for simple tasks wastes money.
Solution: Specify the fast model when appropriate.
{
"model": "enigma-browser-fast",
"taskDetails": "Get the title of this page"
}
4. Optimize Task Descriptions
Problem: Verbose or vague instructions waste tokens.
Solution: Be concise and specific.
❌ Verbose:
“Please navigate to the website and then search for the product I mentioned earlier and then click on the first result and extract all the information you can find about it.”
✅ Optimized:
“Search for [product] and return: name, price, rating from first result.”
Savings: Reduces input tokens by 50-70%.
Problem: Creating separate sessions for related tasks multiplies compute costs.
Solution: Chain tasks in one session.
❌ Expensive:
await runTask("Go to amazon.com"); // New session
await runTask("Search for keyboards"); // Another new session
await runTask("Get first result price"); // Yet another session
const session = await startSession("Go to amazon.com");
await sendTask(session, "Search for keyboards");
await sendTask(session, "Get first result price");
await terminateSession(session);
6. Avoid Unnecessary Page Loads
Problem: Navigating to unnecessary pages burns tokens.
Solution: Be specific about starting URLs.
❌ Wasteful:
{
"taskDetails": "Search Amazon for keyboards"
}
{
"taskDetails": "Search for keyboards",
"startingUrl": "https://amazon.com"
}
7. Use avoidDomains to Prevent Detours
Problem: Agent accidentally navigates to unrelated domains.
Solution: Block known distractions.
{
"taskDetails": "Research this topic",
"avoidDomains": ["facebook.com", "twitter.com", "instagram.com"]
}
Monitoring Costs
Dashboard
View historical usage and costs at app.enigma.click.
API Responses
Every task response includes the usage.cost field for real-time tracking.
Polling Updates
When polling long-running tasks, cost accumulates in real-time:
{
"pending": true,
"usage": {
"cost": 0.0045 // Current cost, will increase as task continues
}
}
Free Tier / Credits
Check your dashboard for current credit balance and any promotional offers.
Billing
- Billing Cycle: Monthly
- Payment Methods: Credit card, ACH
- Minimum Charge: None — pay only for what you use
- Billing Dashboard: app.enigma.click/billing
Cost Comparison
vs. Manual Labor
| Task | Enigma Cost | Human Labor (@ $25/hr) | Savings |
|---|
| Product research | $0.05 | $2.08 (5 min) | 98% |
| Form filling | $0.03 | $1.25 (3 min) | 98% |
| Data extraction | $0.15 | $10.00 (24 min) | 99% |
vs. Traditional Automation
| Factor | Enigma | Selenium/Playwright |
|---|
| Development time | Minutes | Hours/Days |
| Maintenance | Zero | Ongoing |
| Adaptability | Self-healing | Breaks on changes |
| Cost per run | $0.01-0.30 | Infrastructure + dev time |
