Frequently Asked Questions

Register on the developer portal with your email, username and password. You'll automatically get a Free plan with 2,000 requests per month.

Yes. Every new user starts with a Free plan that includes 2,000 requests per month — no credit card required.

Contact our sales team at enquiry@virtisha.com or visit the pricing page on your dashboard.

Yes, you can downgrade at any time. Changes take effect at the start of your next billing cycle.

We accept credit cards (Visa, Mastercard, American Express), PayPal, and bank transfers for Enterprise plans.

We offer a 30-day money-back guarantee for Pro and Enterprise plans. Contact enquiry@virtisha.com.

Every API call that performs face recognition, saves a face, searches the database or updates face information counts as one request.

You'll receive a 429 error response, and the system tells you when your limit resets:

• Hourly limits reset every 60 minutes
• Daily limits reset at midnight UTC
• Monthly limits reset 30 days from your billing date

Yes — contact us to add request packs to your existing plan.

No. Only successful requests (HTTP 200 responses) count toward your limits.

Three ways to keep an eye on your quota:

• Use the /auth/rate-limit-status/ API endpoint
• View usage statistics on your dashboard
• Set up email alerts for usage thresholds

Trained tags are unique faces or identities in your database that the system has learned to recognize.

• JPG/JPEG
• PNG
• Maximum file size: 10 MB
• Recommended resolution: 640x480 or higher

Accuracy depends heavily on image quality:

• Highest on clear, high-quality frontal face images
• Stays strong on moderate-quality images
• Decreases with poor lighting, extreme angles or low resolution

The system detects the first/largest face in the image. For multiple faces, make separate API calls for each face.

• Most requests return quickly
• Response time depends on image size and server load
• Very large databases may take a little longer

We store:

• Face images (encrypted)
• Facial feature vectors (mathematical representations)
• User-provided metadata (names, user IDs)
• Detection logs

Yes. We use:

• AES-256 encryption for stored data
• TLS/SSL for data in transit
• Regular security audits
• GDPR-compliant data handling

Yes. Contact support to request a full data export in JSON or CSV format.

No. Your face data is only used for your account's recognition purposes. We do not use customer data to train our models.

Common reasons:

• Face is too small in the image
• Poor lighting or a very dark image
• Face is at an extreme angle
• Face is obscured (sunglasses, mask, etc.)
• Image quality is too low

Solutions:

• Use well-lit, clear images
• Ensure the face occupies at least 20% of the image
• The face should be looking relatively forward
• Remove sunglasses or face coverings

A number from 0 to 1 (or 0% to 100%) indicating how similar two faces are:

• 0.90-1.00: Excellent match (very likely the same person)
• 0.80-0.89: Good match (probably the same person)
• 0.70-0.79: Moderate match (possibly the same person)
• 0.60-0.69: Low-confidence match
• Below 0.60: Not a reliable match

For best results:

• Minimum: 1 clear frontal image
• Recommended: 3-5 images with different lighting conditions, angles (frontal, slight left, slight right), expressions and time periods (to account for aging)

The system can detect faces with masks, but recognition accuracy drops significantly. We recommend:

• Using pre-mask images in your database
• Training with both masked and unmasked photos
• Considering eye-region recognition features

Yes. The system works with both color and grayscale images. Color images generally provide slightly better results.

Yes, but you need to:

1. Extract frames from the video
2. Send individual frames to the API
3. Process results frame by frame

We recommend processing 1-2 frames per second rather than every frame.

Our REST API works with any language that can make HTTP requests, including:

• Python
• JavaScript / Node.js
• Java
• PHP
• C# / .NET
• Ruby
• Go — and more

Currently we provide API documentation for direct REST access. SDKs are coming soon for:

• Python
• JavaScript
• Java

Yes — you can integrate our API into:

• iOS apps (Swift / Objective-C)
• Android apps (Java / Kotlin)
• React Native apps
• Flutter apps
• Xamarin apps

Not currently — you need to poll the API for results. Webhooks are on our roadmap.

Enterprise customers can request on-premise deployment. Contact enquiry@virtisha.com.

You can make up to 10 concurrent requests. Anything beyond 10 is queued.

Always check HTTP status codes:

response = requests.post(url, json=data)
if response.status_code == 200:
    # Success
    result = response.json()
elif response.status_code == 429:
    # Rate limited - wait and retry
    time.sleep(3600)
elif response.status_code == 401:
    # Bad credentials
    print("Check your API key")

Yes — use the namespaces feature to organize faces into different projects or groups.

Yes. Use the search functionality in the web interface or filter your API calls.

Not currently — you need to make individual API calls. Batch processing is on our roadmap.

Currently you can attach:

• Name
• User ID
• Active status

Custom metadata fields are coming in a future update.

Not currently. This capability is on our roadmap.

No. Our system only does face recognition — identifying who the person is. Demographic or emotion detection is not available.

Problem: invalid credentials.

1. Check your API key and secret for typos
2. Ensure there are no extra spaces
3. Verify you're using the correct credentials
4. Try regenerating your credentials from the dashboard

Problem: missing or invalid data in your request.

1. Check that all required fields are present
2. Verify the image is properly Base64 encoded
3. Ensure the JSON is properly formatted
4. Check the API documentation for correct field names

1. Add more training images per person
2. Use higher-quality images
3. Ensure proper lighting in photos
4. Avoid extreme angles
5. Update old photos if people have aged significantly

1. Check your internet connection
2. Reduce image file size
3. Compress images before uploading
4. Check our status page for service issues
5. Try during off-peak hours

1. Click "Forgot Password" on the login page
2. Enter your email
3. Check your inbox for the reset link
4. Create a new password

Contact enquiry@virtisha.com from your account email. All your data will be permanently deleted within 30 days.

Do:

• Use images of at least 640x480 resolution
• Ensure good, even lighting
• Keep the face at 20-50% of the image
• Keep the face upright and forward-facing
• Use recent photos

Don't:

• Use blurry or grainy images
• Use extreme backlighting or shadows
• Use very small faces in large images
• Use extreme angles (profile views)
• Apply heavy filters or effects

Do:

• Cache frequently used results
• Implement exponential backoff for retries
• Monitor your usage regularly
• Handle errors gracefully
• Use HTTPS for all requests

Don't:

• Store API credentials in your code
• Make unnecessary repeated calls
• Ignore rate-limit warnings
• Process every video frame (sample instead)
• Skip error handling

Do:

• Store credentials in environment variables
• Use secure HTTPS connections
• Rotate credentials periodically
• Monitor for unusual activity
• Implement access logging

Don't:

• Share credentials with others
• Commit credentials to Git
• Use the same credentials across environments
• Ignore security alerts
• Store credentials in plain text

• Email: enquiry@virtisha.com
• Live chat: available on your dashboard (business hours)
• Help docs: the Quick Start and User Guide on this site

• Your username/email
• A detailed description of the issue
• Steps to reproduce the problem
• Error messages (full text)
• Screenshots if applicable
• Your subscription plan

• Email: 24/7 (response within 24 hours)
• Live chat and phone: Monday to Friday, 9 AM - 5 PM
• Emergency: 24/7 for Enterprise customers

Coming soon — we're building a community forum where users can share tips and help each other.