FAQ
FAQ
❓ NV Global KYC Platform FAQ
Below are frequently asked questions about using the platform, its architecture, integration, and the behavior of NV Global KYC modules.
General (KYC)
Question: What is the NV Global KYC platform and what is it used for?
Short answer:It is a comprehensive solution for remote customer identity verification.
Detailed answer
The NV Global KYC platform enables companies to remotely verify user identities using document and face recognition. It automates the KYC (Know Your Customer) process by combining document scanning, biometric verification, and other checks for fast and accurate identity confirmation.
Question: Which document types and countries are supported by the NV Global system?
Short answer:A wide range of documents from different countries is supported — passports, ID cards, international passports, and more.
Detailed answer
The platform recognizes passports, national ID cards, driver’s licenses, and other official documents from many countries. The exact list of supported documents is regularly updated — before integration, it is recommended to verify whether the system supports the documents of your country and the required type.
Question: Can support for a new document type be added if it is not recognized by the system?
Short answer:Yes, if needed, the list of supported documents can be extended.
Detailed answer
If the document you need is not yet supported, you can contact NV Global support. The team can assess the possibility of training the system for this document and releasing an update. Adding a new document type usually requires time to collect data and retrain the model, but it is possible as part of collaboration.
Question: How can we ensure high-quality photos of documents and selfies for successful verification?
Short answer:You need clear photos without glare, shadows, or cropped parts.
Detailed answer
We recommend photographing the document in good lighting, without flash and glare, so that all important areas (photo, stamps, numbers) are fully visible. Selfies should be taken with a good-quality camera, in sufficient lighting, looking directly at the camera with a neutral facial expression — this increases recognition accuracy and reduces the likelihood of errors.
Question: Can scanned copies or photocopies of documents be used instead of photos?
Short answer:Yes, but only if anti-fraud checks for copies and printouts are disabled.
Detailed answer
By default, the platform is optimized for photos of original documents. If anti-fraud checks for photocopies and printouts (both black-and-white and color) are enabled, such images may be recognized as suspicious and flagged with errors like image printed, black and white image, or image edited. However, if necessary, these checks can be disabled in the scenario editor. In this case, the system will successfully accept high-quality copies and scanned documents (including printouts and Xerox copies), provided they meet requirements for resolution, completeness, field visibility, and other parameters. To use copies, make sure the corresponding anti-fraud rules are disabled in the verification scenario.
Question: How long does a full verification (document + face) usually take on the NV Global platform?
Short answer:On production servers — less than a second. With Liveness — up to a minute.
Detailed answer
On NV Global servers, a full document check (including upload, recognition, field extraction, verification, and face matching) takes on average 500–600 milliseconds, and in some cases — even less than half a second. This is achieved thanks to an optimized architecture and high-performance neural networks. On demo servers (in demo mode), the same check may take up to 5 seconds due to limited compute resources and background load. If the scenario includes a Liveness check, the total verification time may increase. In this case, the time it takes a user to complete the Liveness check ranges from a few seconds to one minute, depending on capture quality, lighting conditions, and the user’s reaction speed.
Question: Does the user need access to a camera to complete verification, or can they upload an existing photo?
Short answer:For selfie (and liveness) — a camera is mandatory. Documents can be uploaded, with caveats.
Detailed answer
The user can upload a photo of a document from a file — this is supported by the platform. However, for the face verification stage (including Liveness), the use of a camera is mandatory. This is necessary to confirm liveness and prevent substitution attempts — the system must ensure that a live person is in front of it, not just an image. Nevertheless, for document capture we strongly recommend using the built-in camera component — WebSDK. This significantly reduces fraud risk. If anti-fraud checks in the current scenario are not critical and the scenario allows file uploads without protection against falsifications, then the user can provide document images directly from the gallery or via file upload.
Document Recognition
Question: What image format and size are required for document and face recognition?
Short answer:Only JPEG or PNG. Document — at least 200 pixels; face — at least 30×30 pixels.
Detailed answer
The system accepts images exclusively in JPEG or PNG formats. Files in other formats (for example, HEIC, WEBP, PDF, etc.) will be rejected. The image size requirements refer not to the entire file as a whole, but to the size of the object in the image: The document in the photo must occupy at least 200 pixels along its larger side (width or height). If it is smaller, the system will return a too small image or document too small error because it cannot reliably extract data.
The user’s face (for example, in a selfie or on the document) must be at least 30×30 pixels. Smaller faces cannot be reliably analyzed and recognized and will result in a face too small error.
It is also important to avoid strong compression, blur, and poor lighting — such defects can degrade OCR quality and reduce face matching accuracy.
Question: Do we need to specify the document type or country before recognition, or does the system determine them automatically?
Short answer:The system automatically determines the document type from the image; no separate setting is required.
Detailed answer
NV Global can automatically recognize which document has been uploaded (for example, a passport or driver’s license, and from which country, if the format is unique). You do not need to select the type manually — it is enough to provide a photo and the algorithm will classify the document. However, in some cases (if documents of different types are very similar) the system may require specifying the type in the scenario settings.
Question: What should we do if some document data is recognized incorrectly (an error in name, date, etc.)?
Short answer:It is recommended to retake the document photo with better quality or check the original.
Detailed answer
If the system incorrectly recognizes text on the document, the most likely reason is insufficient image quality (glare, blur). In this situation, it is better to take a new photo of the document with higher quality and try again — in most cases this solves the problem. If the recognition error persists on a clear image, you can report it to support, and for the time being manually correct these data outside the system.
FAQ
❓ NV Global KYC Platform FAQ
Below are frequently asked questions about using the platform, its architecture, integration, and the behavior of NV Global KYC modules.
Document Authenticity & OCR
Question: Does the system verify document authenticity using visual security features (holograms, watermarks)?
Short answer:Indirectly — the system checks structural elements, but direct hologram analysis is not supported.
Detailed answer
The NV Global algorithm analyzes image quality and the presence of mandatory elements (stamps, signatures, holder’s photo) and can detect copies or forgeries. However, full verification of security holograms or UV marks from a regular photograph is difficult. Thus, the system detects obvious inconsistencies (document format, fake elements), but does not guarantee detection of all types of forgeries, especially those that do not clearly appear in the photo.
Question: Does the system support recognition of multi-page documents (for example, passport + registration page)?
Short answer:Yes, for such cases multiple document images can be uploaded sequentially.
Detailed answer
If an identity document consists of two pages (such as a domestic passport with a registration page), the platform can process both pages sequentially. In the scenario settings, you can specify the need for an additional step — first the main page, then, for example, the registration page. The user will be asked to upload both sides/pages, and the system will combine this data for a complete check.
Question: Does the system recognize text on documents in different languages (for example, Cyrillic, Latin)?
Short answer:Yes, OCR supports many languages, including Cyrillic, Latin, and others.
Detailed answer
NVGlobal can read text in a wide range of languages, including those that use Latin-based scripts (English, French, etc.) and many other writing systems commonly found in identity documents. The system is trained on international samples and automatically extracts fields (full name, date of birth, document number, etc.) regardless of the language — it is only important that the document is in the list of supported formats.
Liveness (Proof of Life)
Question: How does the user liveness check (Liveness) work?
Short answer:The user performs simple actions on camera, and the system determines whether a live person is in front of it.
Detailed answer
The Liveness check on the Enface platform is used to confirm that a real live person is in front of the camera, and not a photo, video, or other spoof. We use three modes of active Liveness:
Simple tasks — the user performs individual commands (for example, turn their head left or right, blink, look at a certain point).
Complex tasks — tasks are more sophisticated and follow in a quick sequence (for example, combinations of head turns and gaze directions) for deeper verification.
Circular head movement — the user moves their head in a circle, similar to the initial Face ID setup on an iPhone. This method allows collecting a maximum amount of 3D information about the face.
While the tasks are being performed, the system captures a series of keyframes that are automatically sent to the server. On the server, these images are analyzed using neural network models that determine signs of a live person (micro-movements, skin texture, depth, etc.). After analyzing all frames, the system issues a final decision on whether the Liveness check has been passed.
Question: What should we do if a client fails the Liveness check (live presence could not be confirmed)?
Short answer:The client should try again following the instructions — most repeat tests pass successfully.
Detailed answer
If the Liveness check is not passed, the system could not confirm that a live person is in front of the camera (for example, due to poor lighting or failure to perform the required actions). It is recommended to ask the user to repeat the Liveness check: look directly into the camera, not cover their face, and complete the requested action. The system usually allows several attempts. If it still fails, you may switch to an alternative identification method (for example, manual review by an operator).
Question: Can the Liveness step be disabled so that only document verification is performed?
Short answer:Yes, it is possible to configure a scenario without Liveness, although this reduces the level of protection against impersonation and fraud.
Detailed answer
The Liveness test is recommended to ensure that it is the actual document holder presenting the document at the moment of verification. However, the platform is flexible — at the client’s request, you can use a simplified KYC scenario that includes only document upload and recognition. In this case, the user’s face is not confirmed through a selfie, which slightly increases fraud risk, so disabling Liveness is reasonable only in low-risk scenarios.
Question: Does the system detect spoofing attempts during Liveness (masks, someone else’s video instead of a real face)?
Short answer:Yes, Liveness algorithms can detect inconsistencies (for example, a mask on the face or a video feed).
Detailed answer
The built-in Liveness mechanisms check the naturalness of behavior: they can distinguish a live face from a static photo or a recorded video. If, for example, an attacker puts on a silicone mask or tries to show a screen with a video of another person to the camera, the system will, with high probability, recognize this as a spoofing attempt and will not count the Liveness check as passed.
Anti-fraud
Question: How does the platform detect fraudulent attempts during identification?
Short answer:It analyzes document forgery signs, repeated use of biometrics, and other anomalies.
Detailed answer
NV Global implements a number of anti-fraud mechanisms: it checks the document for forgery (for example, detecting a copy or photomontage), compares a person’s biometrics with those already known to the system (to detect whether the same person tried to register under another name), and monitors the sequence of user actions. Any anomalies (black-and-white document copy, mismatch between face and data, suspiciously frequent attempts, etc.) are flagged and can serve as a signal for rejection or additional manual review.
Question: What does the “[fraud] image printed” error mean during document verification?
Short answer:It is a warning that the uploaded document image looks like a photocopy or printout.
Detailed answer
The [fraud] image printed label indicates that copy-related signs were detected: the system determined that the document is likely not an original, but a scan or photocopy (for example, black-and-white or color). This error is a signal of possible fraud. It is recommended to request a photo of the original document — with real security features — for the user to pass the check.
Question: Can the system detect that the same person is trying to pass identification under different names?
Short answer:Yes, the system detects face matches and links them across sessions.
Detailed answer
The platform supports storing biometric face templates of users who have successfully passed verification. During new KYC attempts, the system automatically compares the face with the database of existing embeddings.
If a face has already appeared in previous sessions — even under a different name, with a different document, or in another context — the system will detect a match. This allows you to promptly identify repeated KYC attempts with substituted personal data.
In the interface, when a match is found, the ID of the current session is highlighted in red, which is a clear indicator that this person has already appeared in the system. By clicking this ID, you can open the history of all sessions in which this face was detected, with the ability to review and analyze them.
Question: Can I configure a “stop-list” – a list of unwanted persons or documents for blocking?
Short answer:Yes, we have implemented a face-based blacklist mechanism.
Detailed answer
The platform supports a built-in mechanism for adding faces to a blacklist. This allows automatically blocking the verification process for the same person, even if they use other documents or personal data.
To do this, in the session view interface you need to click on the icon with a crossed-out circle. After that, the face detected in this session will be added to the blacklist. For subsequent verification attempts:
If the face matches a previously blocked one — the session will be automatically rejected;
In the interface this will be accompanied by the label “User is blacklisted”.
Blacklist in the interface
This approach is effective against repeated fraudulent attempts with data substitution, as it relies on biometrics rather than only document details. The mechanism does not require any additional implementation on the client side — it works out of the box.
API integration
Question: Where can I find the NV Global API documentation for integration? Short answer:The official website provides detailed API documentation (including a Swagger interface).
Detailed answer
All technical documentation is located on the NV Global portal: it describes the REST API endpoints, request and response formats, examples, and errors. For easier integration, you can use Swagger/UI at https://api.nv.global/docs/ or read the guide on GitBook. Before starting integration, it is recommended to review these materials.
Question: How do I get access to the API (test and production keys)? Short answer:You need to request a promo code from the administrator and register on the platform.
Detailed answer
To gain access to the platform’s API, you need to:
Request a unique promo code from a NV Global administrator.
Register on the platform using this promo code (via the web interface).
After registration, you will gain access to the personal dashboard, where you can:
All access parameters (tokens, identifiers, configurations) are issued automatically through the platform — we do not manually provide anything, including any clientKey or secret keys. Usually, a confidentiality agreement (NDA) is concluded beforehand. After this, you can start using test access. Load parameters, limits, scenario restrictions — are discussed individually. After concluding a full contract, production access is issued; connection to an isolated environment, dedicated servers, and custom SLA parameters are possible.
Question: What are the main steps of API integration for KYC verification?
Short answer:You need to send photos to the server using the appropriate API method and receive a verification result.
Detailed answer
The general process is as follows: your backend calls the NV Global API method, sending the document image and the user's selfie/video (either with sequential requests or within a single session, depending on the chosen scheme). The NV Global server processes the data (OCR, face matching, liveness) and returns a result structure — the verification status and extracted data. Your integration must implement sending data, receiving responses, and parsing results (for example, filling in the user's profile with passport data, handling the “success/failure” status).
Question: Are the document and selfie sent in one request or separately?
Short answer:They can be sent in a single request or step-by-step — synchronously or asynchronously.
Detailed answer
The NV Global platform supports different integration schemes — both using the ready-made widget and through direct API requests.
If you use the web widget, everything happens automatically: it sends all data (document, selfie, liveness) step-by-step and in the correct sequence, creating a single session.
For direct API integration, you can choose between several approaches:
In a single request — some scenarios allow sending both the document and the face at once (e.g., with a static selfie holding the document).
Step-by-step — the more common sequential scheme:
A session is initiated,
The document photo is sent,
Separately — the selfie.
Both synchronous and asynchronous processing modes are supported: you can get the result immediately or via webhook/polling.
This approach provides maximum flexibility — you can optimize UX, adapt to specific scenarios (e.g., document-only, selfie-only, selfie with document, liveness, etc.), and fully control the verification logic.
Question: What is sessionId and how do I use it?
Short answer:sessionId is the unique identifier of the verification session used to track the KYC process.
Detailed answer
When the verification is initiated, the system assigns a unique sessionId to the request. You need to use it in subsequent requests: for example, to link a selfie to an already submitted document, or to request the verification status/result. The sessionId helps correlate responses with a specific identification attempt, especially if you have parallel verifications running for different users.
Question: Is it possible to reuse sessionId for another client or a new verification?
Short answer:No, a separate unique session (sessionId) is created for each individual verification.
Detailed answer
SessionId is uniquely linked to a specific KYC attempt. If one user fails and the verification starts again (or another user goes through KYC), a new sessionId is generated. Reusing an old identifier is not allowed – this will lead to data mixing. Thus, each verification has its own session, and completed or expired sessionIds are not used in the future.
Question: How can I get the verification result: does the platform send the response itself or do I need to poll the status?
Short answer:By default, you can poll the session state; a callback (webhook) for the final result is also available.
Detailed answer
You can periodically request the KYC session status by sessionId until you get a final response (SUCCESS/FAILED). However, it is more convenient to configure a callback URL: then, after the verification is completed, NV Global will itself send a notification with the result to your server. The webhook notification will contain the sessionId and the final status (and data), which eliminates the need for polling. The way the result is received is configured during integration.
Question: Is there a reference of all possible errors and codes returned by the system?
Short answer:Yes, it is available in the KYC errors section and contains a detailed description of each check.
Detailed answer
The full list of all errors and codes that the system can return during the KYC verification process is available in a special reference section — the “KYC errors” section.
The reference contains:
Check code (for example,
[general] bad image,[fraud] mobile device);Error description — what exactly happened;
Comment — how to interpret it from the user’s point of view;
Technology description — which algorithms were triggered and why the error occurred.
Examples from the reference
[general] bad image— the file is not an image (for example, not JPEG/PNG or corrupted);[document] text fields not visible— key document fields were not read;[fraud] mobile device— the document or selfie was taken from a smartphone screen.
The reference is presented in tabular form and is regularly updated. It is used both by technical specialists and support operators.
If during integration or verification an unknown code appears that is not in the reference, you can always contact support — we will provide its explanation and recommendations on how to handle it.
Question: Does the platform’s API return recognized personal data (full name, date of birth, etc.) in a structured form? {{ ... }}
Short answer:Yes, all data is returned in structured JSON format.
Detailed answer
If the document is successfully recognized, the platform returns personal data as a structured JSON where each field contains:
Title (title) — for example, Surname, Date of Birth, Document Number;
Value (value);
Confidence level (conf);
Validity status (for example, isMismatch, isUnreadable, isEdited);
Image indexing (imageIndex), if there are multiple images.
The format is universal and suitable for both international documents (e.g., ePassports, ID cards) and national identity documents of various countries. Example of a structure for a foreign passport:
Similarly, for a UK passport:
Additionally, the response may include:
Machine Readable Zone (MRZ Strings);
Date fields in ISO format (valueUTC);
Specialized fields such as Issuing authority, Nationality, Place of Birth;
Metadata on authenticity checks (markers isPrinted, isEdited, isXerocopy, isMobile, etc.).
The JSON response also contains blocks with fraud checks and face matching results, with similarity percentages between the document, selfie, and selfie with the document.
Question: Why can the verification result be FAILED even though the data was recognized correctly (for example, all passport fields were read correctly)?
Short answer:This means the system is not confident in the correctness — for example, the model’s confidence threshold was triggered despite the apparent recognition.
Detailed answer
In such cases, the final status “FAILED” indicates that the identification algorithm considered the verification unsuccessful. For example, the neural network might have recognized the text but assessed its reliability as low (low confidence) — in this case, the system prefers to refuse, even if the data looks correct. Another example is when the user’s face matches the photo, but Liveness is not passed, or the document is read but deemed invalid. Thus, even with filled data, the platform can return a refusal if at least one of the verification criteria is not met.
Web widget and SDK
Question: Is there a ready-made web widget for passing KYC that we can embed on our site?
Short answer:Yes, NV Global provides the “KYC Widget” for simplified integration on websites.
Detailed answer
Instead of developing your own interface, you can use the ready-made widget from NV Global. It is an embedded component that shows the user all necessary steps (passport upload, selfie, liveness) in a clear way. The widget is easy to connect to your web page with a few lines of code and automatically interacts with the API – you do not need to implement the data exchange logic with the server yourself.
Question: How do I integrate the KYC widget into my web page or application (for example, a React application)?
Short answer:You need to connect the NV Global widget script and call initialization with your schema identifier (schemaId).
Detailed answer
The process is very simple: you add the widget JS library to the page (``). Then you call the window.KYCWidget.setupKYC({...}) method, specifying the schemaId of your verification and clientKey (if required). In React or other frameworks, you can install the kyc-widget package from npm. The documentation provides examples: it is enough to call the widget opening method on a button click — everything else happens automatically.
Question: Is there a mobile SDK for integrating NV Global into our Android/iOS application?
Short answer:There is no specialized native SDK yet, but you can embed the web widget into a mobile application.
Detailed answer
Currently, NV Global offers a web widget that can also be used in mobile applications (for example, via WebView or an embedded browser inside the application). Separate libraries for Android/iOS are not provided yet, however the web solution is adaptive and supports smartphone cameras. Thus, even without a native SDK you can implement the KYC process on mobile platforms using the existing web component.
Question: Is it possible to change the appearance of the widget to match our brand (colors, logo)?
Short answer:The widget can be branded by adding your logo and changing the color scheme via styles.
Detailed answer
The widget interface is neutral and can be integrated into your site as part of the page. You can place your company logo next to it or in the header, and the widget also supports some style settings (for example, you can override the CSS to change the main colors to match your brand). Full white-label customization of the interface is limited by the widget’s capabilities, but basic appearance customization is possible so that the KYC form fits organically into your product.
Question: What is schemaId in the context of the widget and where can it be obtained?
Short answer:schemaId is the identifier of the verification scenario configured for you in the NV Global system.
Detailed answer
The platform can have several KYC schemes (scenarios) – for example, a simplified check of only the passport, a full passport+selfie check, a check with additional documents, etc. Each scheme is assigned a unique schemaId. When connecting the widget, you need to specify the schemaId that corresponds to the desired process. This identifier will be provided to you by the NV Global integration manager, or you can see it in your personal account when configuring the system.
Question: What is clientKey and how can it be obtained for the widget or API to work?
Short answer:ClientKey is a client access key issued by the system for request authentication; it can be obtained from support or in your personal account.
Detailed answer
When integrating the widget or direct API requests, clientKey can be used to confirm your access rights. This key is linked to your account and is issued by NV Global for your applications. It is usually obtained together with other integration data (such as test and production keys). In some cases the widget can work without explicitly specifying clientKey (schemaId is sufficient if the key is bound on the server side). Always follow the documentation instructions regarding the use of clientKey.
Errors and codes
Question: What does the error "Required elements are missing from the document" mean in the verification result?
Short answer:It means that some mandatory fragments are not visible on the photo of the document (a stamp, signature, or other important elements).
Detailed answer
This error text appears when the system has not detected all the required areas on the image of the document. For example, the owner's signature, the issuing authority's stamp, or other mandatory requisites of the passport/ID card may be missing. In practice this error is often caused by the fact that part of the document is out of frame or obscured. It is recommended to take a new photo that shows the entire document completely; then the verification will be successful if all elements are present.
Question: What does the "bad_image" error mean when uploading a photo?
Short answer:The system failed to process the file — it may not be an image or it may be a low-quality image.
Detailed answer
The bad_image error means that the uploaded file cannot be processed as an image.
Main reasons for the error:
The file is not an image at all (for example, PDF, TXT, DOCX or another unsupported extension);
The image format is not supported — only JPEG and PNG are allowed;
The file is corrupted or encoded with errors;
In rare cases, the image is so blurred, dark, or distorted that the system cannot start processing.
What to do to fix the error?
Make sure that you are uploading an image in JPEG or PNG format;
The file size is not zero and it is not corrupted;
The document or face is clearly visible in the image, without blur, skew, or excessive darkness.
If the file is valid but the error persists, it is recommended to photograph the object again and repeat the upload.
Question: What does the error "bad_face_matching" mean in the verification result?
Short answer: It means that the face in the selfie did not match the photo in the document with sufficient confidence.
Detailed answer
The bad_face_matching error indicates an unsuccessful biometric verification:
The system compared the user's face image with the photo in the document and concluded that these are different people (or the similarity score is too low).
In other words, the algorithm did not confirm that the selfie belongs to the document holder.
Possible reasons for the error:
A completely different person;
Poor photo quality (blur, shadows, angle, lighting);
Strong differences in appearance between the document photo and the selfie.
Recommendation:
If the user is sure that everything is correct, it is worth trying again with other photos — the previous ones may have been blurry or taken from a different angle.
Question: What result statuses can the system return and what do they mean?
Short answer: The main statuses are — success, failed, idle, and suspicious is also possible.
Detailed answer
The system returns several key statuses as a result of identity and document verification. Each status reflects the current state or the final verification result:
success — verification has been successfully completed. All verification steps are completed, the data is confirmed, the face matches the document, no errors are found.
failed — verification has not been passed. One or more verification steps resulted in an error (for example, the document is invalid, the face does not match, or signs of falsification were detected).
idle — the session remained unfinished. The user did not complete the process: did not upload a selfie, closed the window before finishing, did not pass liveness, etc.
suspicious — the session is marked as suspicious. This may happen if:
The user is found in the blacklist;
A repeated face with different data is detected (for example, one face — two different passports);
During a single session, attempts were made to upload several different documents with different faces (a sign of fraud).
Additionally, the API response structure may contain intermediate statuses of components (for example, ocr, fraud, face_matching) — they allow you to analyze at which stage errors occurred, but the final session status will be one of the four listed above.
Question: What does the "idle" status mean and in which situations is it set?
Short answer: "Idle" means that the verification was not completed by the user.
Detailed answer
The idle status appears if the user started the process but did not provide all the required data. \
Typical situations when idle occurs:
The user uploaded the document but did not take a selfie;
The Liveness step was not completed;
The user closed the window or left the page before completing all steps;
The user did not confirm or did not submit the required information.
What it means for integration:
The final verification did not take place — the status is neither success nor error;
Such sessions are usually considered abandoned and closed after a timeout;
You can decide for yourself how long to wait for the user and what to do with such sessions (for example, notify the user or delete the session).
Question: How can I understand the reason why the verification was not passed (status failed)?
Short answer: The reasons are specified in the results array, in the errors list for each verification type.
Detailed answer
If the KYC platform response contains the status failed, this means that one or several parts of the verification were not successfully completed. \
Where to look for the reason for the rejection:
The results array — contains separate objects for verification types (document, selfie, faces, etc.).
Each element has an errors field — a list of error codes.
This means:
It was detected that the document or selfie was photographed from a mobile device screen ([fraud] mobile device);
Important document fields, such as place of birth, were not recognized on the image ([document] text fields not visible).
Checking antifraud results (checks):
This confirms that the mobile device check (isMobile) failed with high confidence (0.947).
Document fields and readability status (ocr):
This means that the "Place of birth" field could not be recognized, which can also be a critical error depending on the scenario.
Step by step, to understand the reason for the rejection:
Check the status field at the top level (
"status": "failed").Review the nested results — for each type (document, selfie, etc.).
Look at the errors field inside these blocks — it contains the main reasons.
Check checks and ocr if you need to understand the details (for example, which fields are unreadable or what caused the suspicion).
On-Premise vs Cloud
Question: Is it possible to deploy the NV Global platform on our own servers (On-Premise) instead of using the cloud service?
Short answer: Yes, an On-Premise version with installation in the client's infrastructure is available.
Detailed answer
For companies with increased requirements for security and data confidentiality, the NV Global platform provides an On-Premise solution — a software suite for local deployment within the client's IT infrastructure. \
In this option:
All images and results are processed and stored only on the client side, without going to external cloud services;
The system is delivered as a set of Docker containers ready for installation;
Isolated environments are supported, including complete absence of internet access.
Installation process:
NV Global specialists connect to the client's servers (usually via SSH or through a secure channel);
They perform the installation of all components, environment configuration, and operability checks;
Later, if updates or technical support are required, the NV Global team can remotely connect to the server and carry out upgrades, model updates, or issue resolution.
This solution is suitable for banks, government institutions, medical organizations, and other entities where data control requirements are critically important.
Question: What is the difference between the cloud version of the service and the On-Premise version?
Short answer: The cloud version runs on NV Global servers (SaaS), while On-Premise runs on your servers, giving full control over data.
Detailed answer
In the cloud option, the entire infrastructure is maintained by the provider: you only need to use the API over the internet, while updates and scaling are handled by NV Global. \
In the On-Premise option, you deploy the system on your side: this provides maximum autonomy and compliance with security policies, but requires resources for installation, maintenance, and updates.
Key differences:
Hosting location: Cloud version — on NV Global servers, On-Premise — on your servers.
Data control: On-Premise provides full control and isolation of data.
Updates and support: In the cloud this is handled by NV Global, On-Premise requires independent or joint maintenance.
Functionality: The platform core is the same; the differences are in the place of execution and responsibility for maintenance.
Question: What are the minimum hardware requirements for installing the On-Premise version?
Short answer: At minimum — RTX 3090 or an equivalent with 24 GB VRAM, 16-core CPU, 64 GB RAM. It is recommended to use two servers with load balancing.
Detailed answer
For an On-Premise installation of the NV Global platform, high-performance hardware is required, capable of processing images and biometrics in real time. Below are the minimum and recommended system requirements. \
✅ Minimum requirements:
CPU: AMD Ryzen 9 3900X (or an equivalent with ≥16 threads and AVX2 support)
GPU: NVIDIA RTX 3090 with 24 GB of video memory, or a comparable NVIDIA card in terms of volume and performance (for example, A5000, RTX 4080)
RAM: 64 GB
Disk: SSD from 1 TB
This minimum configuration ensures comfortable operation under moderate load and a small number of parallel sessions. \
✅ Recommended requirements:
CPU: Intel Core i9-12900KF or an equivalent with ≥24 threads
GPU: 2× NVIDIA RTX 3090 (total 48 GB VRAM) or comparable cards — for example, 2× A5000, A6000, RTX 4090
RAM: 128 GB
Disk: SSD/NVMe from 2 TB
Network: 1 Gbit/s and higher
This configuration allows the system to operate under high load with many parallel requests. \
🔁 Redundancy and fault tolerance:
For production environments with high SLAs, it is recommended to use two servers:
A load balancer is configured for both servers, which provides stability in case of failures and scalability.
Question: How are updates to the On-Premise version carried out (new models, improvements)?
Short answer: Updates are performed by the NV Global team via a secure connection to the client's server.
Detailed answer
Updates to the On-Premise version of the NV Global platform include:
new versions of neural network models (OCR, antifraud, Face Matching, Liveness, etc.);
updates of infrastructure components (Docker images, API);
bug fixes and performance improvements.
The update process is organized as follows:
NV Global notifies the client about the available update and agrees on the time for its rollout.
NV Global specialists connect to the client's server via a secure channel (for example, via VPN or according to an agreed corporate procedure).
They perform:
updating of all necessary Docker images and components;
full testing on the client side after the update;
verification of the correct operation of key functions.
After confirmation, the new version is considered active, and all processes switch over to it.
Question: Can the NV Global platform be deployed on your own servers (On-Premise) instead of using the cloud service?
Short answer: Yes, an On-Premise version is available for installation within the client's infrastructure.
Extended answer
After you deploy the NV Global platform inside your network, all checks happen locally. \
Features of full isolation:
No internet connection is required for the algorithms to work or for data storage — everything stays within your organization’s perimeter.
All image and biometric processing takes place on your servers.
Data never leaves the client's infrastructure.
When the internet may be needed:
License activation;
Downloading updates (can be performed in a controlled manner, e.g., temporarily allowing access or using offline packages).
Thus, the On-Premise option is suitable for fully isolated environments — for example, in banking or government systems requiring maximum autonomy and data protection.
Digital Passport (Digital ID)
Question: What is meant by the “Digital Passport on blockchain” mentioned by NV Global ?
Short answer: It is a concept of a decentralized digital identifier stored in a blockchain network for reusing verified data.
Extended answer
NV Global is developing a Digital ID solution, in which a record (a digital passport) is created in a distributed system (blockchain) based on the completed KYC.
What a blockchain-based digital passport is:
A unified verified personal profile created after successful KYC;
Stored in a decentralized and immutable blockchain network;
Allows reuse of verified data across multiple services without repeated identification;
Can be implemented as a Soulbound NFT token or another digital certificate;
Recognized by ecosystem partners, simplifying access to various services.
In essence, this is a layer on top of KYC: complete identification once — receive a digital certificate to present again.
Question: Can the decentralized digital passport from NV Global already be used?
Short answer: This functionality is in the deployment phase — for specific use, details must be discussed with the NV Global team.
Extended answer
The “digital passport” project is relatively new, and its integration is being conducted in cooperation with interested clients and partners. \
Current status and capabilities:
Digital ID functionality is implemented upon agreement with clients;
To connect to the decentralized identity ecosystem, you need to contact the NV Global team;
In the future, the system will become publicly available — users will be able to use a unified digital identity across organizations.
Pricing and Limits
Question: What pricing model does NV Global use — per number of requests, subscription-based, or another?
Short answer: Various models are available — typically either pay-per-use or subscription plans.
Extended answer
The service uses flexible pricing:
For small volumes — pay-per-use (e.g., a fixed price per verification);
For large clients — subscription plans with a prepaid number of checks per month or year;
Pricing and cooperation models are defined individually in the contract.
Thus, NV Global can offer both pay-per-use and fixed-volume packages depending on your needs.
Question: Are there limits on the number of checks per month under our plan?
Short answer: No, you can use your purchased verification package at any pace — even in a single minute.
Extended answer
The NV Global platform operates on a package model: you purchase a specific number of KYC checks, which can be used at any time and at any speed, with no monthly, daily, or hourly limits. \
For example:
If you purchase a package of 10,000 checks, you may use them all in one day or distribute them across a year — there are no time limits.
The platform does not restrict request frequency as long as your infrastructure can handle the load.
Thus, all limits are quantitative, not time-based.
Question: What happens if we exceed the purchased number of checks?
Short answer: The service will not stop — you simply need to buy an additional package.
Extended answer
If you exhaust your purchased package of checks, nothing critical happens: the system will continue processing requests normally. \
You will need to purchase an additional package corresponding to your expected volume.
There are no automatic blocks or shutdowns — this behavior is intentionally avoided to prevent impact on your business processes.
You can track usage in advance and receive notifications from your internal accounting system or NV Global account manager.
Thus, exceeding the package simply means it's time to buy the next volume, not a risk for service stability.
Question: How can we get a report on the number of KYC checks we have performed for a given period?
Short answer: You can generate a detailed report for any period directly in your dashboard.
Detailed answer
To obtain a report on the number of KYC checks performed for the required period, simply use the "KYC Reports" tab in the NV Global platform dashboard. \
In this section you can:
select any date range;
filter sessions by status (success, error, suspicious, not completed);
filter by error type or field values;
see the total number of sessions found;
download the report in CSV or Excel format — it will include all key data: time, status, user name, document type, error (if any), etc.
This allows you to independently track activity, reconcile statistics with your tariff package, and analyze the quality and correctness of passing the checks. Additionally, NV Global generates an official report with the number of checks every month, which is sent to you for approval and invoicing. If necessary, you can request extended details — with sessionId, results, and anomaly indicators for each session.
Question: Do you provide closing documents and service reports for the services rendered every month?
Short answer: Yes, monthly service reports are generated with the number of checks performed, which are provided to the client for signing.
Detailed answer
Under the NV Global contract, at the end of each reporting period, a package of closing documents is prepared. \
Typically this includes a service acceptance report specifying how many verifications were performed during the month (or other period) according to the accounting system.
These reports are sent to your authorized representative (for example, the accounting department) for signing.
You will have official documents confirming the volume of KYC services rendered for settlements and reporting.
Question: Can we switch to an annual plan with a larger volume at a discount after the trial period?
Short answer: Yes, NV Global is ready to discuss switching to an annual service package on favorable terms after a successful trial period.
Detailed answer
Many clients first perform integration and pilot use on a monthly basis to make sure it is effective. \
After that, a long-term contract is usually concluded — for example, for a year — with fixed terms (often a lower cost per check for large volumes).
If you need a predictable annual plan, you can agree on it with your NV Global account manager.
As a rule, this provides a price advantage compared to month‑to‑month billing.
Sessions and statuses
Question: How long is an unfinished session (idle) kept, and do we need to close it manually?
Short answer: A session automatically expires according to the scenario’s timer — you do not need to close it manually.
Detailed answer
Each KYC session has a lifetime that is configured in the scenario settings. \
If a user starts the process but does not complete it (for example, uploads a document but does not pass the selfie or Liveness), the session moves to the idle (not completed) state.
After the configured lifetime expires:
the session automatically receives the status expired;
no manual closing is required on your side;
such sessions are stored in the system and can be viewed in reports as incomplete.
In your business logic, you can set your own timeout for waiting on the user (for example, 15 minutes), and if it is exceeded — suggest that they start the process again or initiate a new session via the API. Thus, managing the session activity period is part of the scenario builder, and its duration can easily be adjusted to match your specific processes.
Question: Do we need to store the sessionId and results ourselves, or can we retrieve the data from the system again later?
Short answer: Yes, you can retrieve all data by sessionId at any time. Images — within 6 months.
Detailed answer
The NV Global platform stores all KYC verification results indefinitely, including:
the session status (success, failed, idle, suspicious, expired);
recognized document fields (OCR);
face matching results;
metadata, errors, fraud checkers, and more.
To access the results at any time, it is enough to know the sessionId:
retrieve all data via the API;
view results and images in the dashboard;
export a report to CSV/Excel.
Original images (documents, selfies, Liveness) are stored for 6 months. After this period the images are automatically deleted or overwritten, but the remaining structured data remains available without limitations. For business convenience, you can store key fields on your side (for example, full name, date of birth, status), but this is not required in terms of access to history — everything can be retrieved again via the API.
Post-Deployment
Question: How can we contact NVGlobal support if issues or questions arise after launch?
Short answer: Support is available during business hours. 24/7 is available. Requests are accepted via tickets, Telegram, and email.
Detailed answer
Immediately after launching the integration, you retain access to NVGlobal technical support. \
Support is divided into two levels:
🔹 Basic support (included by default):
Schedule: 5 days a week, from 10:00 to 18:00 GMT
Communication channels:
Ticket system in the personal account — the main and preferred contact method;
Telegram chat — for informal or urgent questions (as available);
Email support — for general or technical inquiries.
🔹 Extended support (per contract):
Schedule: 24/7, including weekends and holidays;
Enabled separately, per SLA: priority line, faster response, and monitoring.
📌 Important: in many cases, answers are already available:
In the documentation accessible from the personal account;
Through the bot built into the platform or Telegram;
In the FAQ and error reference guides.
If you still cannot find the answer, create a ticket describing the issue and, if possible, include the sessionId, your organization name or scenario — this will speed up processing.
Question: Can we change the verification scheme settings or add new documents after the service is launched?
Short answer: Yes, everything is editable through the KYC editor — you can change parameters, steps, document types, and rules.
Detailed answer
After launching the KYC service, you retain full control over its logic and composition. This is done via the built-in scenario editor (KYC Editor), available under KYC/AML → KYC Editor. \
Through the editor you can:
Create and edit scenarios without developers;
Configure:
secret key (if callback is used);
client list (if known in advance);
blacklist of faces;
maximum number of critical errors before blocking the session;
session lifetime until freeze (e.g., 15 minutes);
data retention period (e.g., 0 — infinite);
Enable anti-fraud settings:
blocking when two different documents are uploaded in one session;
blocking when the same person submits documents with different data;
Configure age filter by passport date of birth (e.g., 16–100 years);
Specify callback URL for backend notifications, with encryption, PII filtering, and rule configuration.
Verification steps (Steps):
Each scenario consists of one or more verification steps:
Step 1 — e.g., uploading a document (national passport, international passport, etc.);
choose required countries and document types;
enable checks: MRZ, editing traces, screen detection, mobile device detection, xerocopy, cutouts, logic checks, etc.
Step 2 — e.g., selfie or selfie with document;
enable checks: face visibility, age, gender, editing traces, screen, mobile device, face matching.
Changes in the scenario take effect immediately after saving — you can adapt KYC to any business logic: from simple to multi-level flows.
Question: What should we do if the service is slow or unavailable?
Short answer: Contact support — we will check during business hours. Outages are rare and usually fixed within minutes.
Detailed answer
If you notice platform slowdown or API unavailability, please report it to technical support via ticket, email, or Telegram. \
⏱ Basic technical support (default) operates on weekdays during business hours (Moscow time). We respond promptly to all requests, including diagnostics and recovery.
📌 Important: 24/7 extended mode is available only with a separate agreement.
What happens during an outage:
An automatic monitoring system tracks the health of all cloud components;
In case of failure, engineers receive an alert and begin resolving the issue without waiting for client requests;
Most outages (when they occur at all) are resolved within minutes — and we try to notify clients in advance if it affects their processes.
For On-Premise solutions, diagnostics can be performed together with your IT team — our engineers can connect remotely if needed and help restore availability.
If you have additional questions — write to us and we’ll add them to the FAQ!
Last updated