MetaDefender Cloud is OPSWAT’s cloud file-security platform. Depending on the workflow and license, it can combine hash reputation, multiscanning, Deep Content Disarm and Reconstruction, sandbox analysis, file-type checks, vulnerability information, and data-loss prevention through a web interface and APIs.
The platform is not one fixed “30-engine scanner.” Community results, commercial API tiers, private modes, and on-premises MetaDefender products have different engines, limits, retention, and sharing behavior. Those distinctions matter more than a generic feature checklist.
This review checks OPSWAT’s current product page, cloud documentation, release notes, privacy guidance, and API licensing information as of August 30, 2026. Top5soft did not upload malware, benchmark detection or speed, purchase a license, or test Deep CDR output for this edition.
Correction published August 30, 2026: The previous edition published conflicting free quotas, unsupported per-file and on-premises prices, fixed speed and detection claims, and generated interface images. It also mixed community engine counts with commercial tiers and overstated privacy. This edition removes those claims and documents standard scanning, Private Scanning, and Private Processing separately.
What MetaDefender Cloud is
MetaDefender Cloud is designed to inspect files entering web applications, cloud storage, SaaS services, and other content workflows. OPSWAT’s current product page lists Metascan Multiscanning, Deep CDR, Proactive DLP, sandboxing, reputation, metadata extraction, and APIs as separate technologies that can be combined in a policy.
That modular model is important. A hash lookup does not upload file content. Multiscanning asks multiple anti-malware engines for signals. Deep CDR rebuilds supported files while removing or changing active and out-of-policy content. A sandbox observes supported files in an execution environment. DLP identifies or redacts sensitive data. One result does not imply that every layer ran.
The community web service is useful for personal evaluation and limited research. Commercial Threat Prevention API tiers are intended for production integrations. MetaDefender Core and related products can run in customer-controlled environments, including air-gapped networks, but they are not the same deployment as the public cloud website.
Standard, private, and private-processing modes
Standard MetaDefender Cloud submissions participate in OPSWAT’s threat-intelligence ecosystem. The official privacy documentation says users give OPSWAT permission to share submission results with the cybersecurity community and to share submitted executables with malware-exchange partners. It also tells users not to submit personal data.
Paid customers can enable Private Scanning. In that mode, the file is held temporarily for processing, removed after analysis, and not stored or shared. Scan results remain in the MetaDefender Cloud database. This protects the file differently from the report metadata.
Private Processing adds another control. OPSWAT says processing results are available only to the owner using the submitting API key, are not added to the shared hash collection, and cannot be retrieved through ordinary hash lookup. This can cover multiscanning, sandbox, and DLP results. It is a paid capability and must be explicitly selected or enforced by organization policy.
When Deep CDR is used privately, the sanitized output remains downloadable for up to 24 hours unless it is deleted earlier. “Private” therefore does not mean no cloud storage at any moment; it means temporary file handling and restricted or non-shared results according to the selected mode. Verify the exact contract, organization policy, region, and retention controls before sending sensitive content.
Start with reputation and hash lookup
The cloud homepage documentation says file analysis starts with a hash lookup. Users can also search directly by MD5, SHA-1, or SHA-256, as well as look up URLs, domains, and IP addresses.
For a confidential or large file, calculating SHA-256 locally and searching first can avoid an unnecessary upload when an existing report already answers the question. Hash lookup reveals the identifier but not the file contents. It works only for the exact file: changing one byte, rebuilding an executable, or repacking an archive changes the cryptographic hash.
An existing reputation result is evidence, not a verdict. Known-good reputation can reduce repeated analysis in a controlled policy, while known-malicious reputation can support blocking. “Unknown” means the database lacks a conclusive reputation for that hash; it does not mean safe.
Before uploading, classify the file and choose standard, Private Scanning, or Private Processing deliberately. Do not rely on account login alone to make a standard submission private, and do not assume that a browser extension changes the cloud account’s license or sharing policy.
Multiscanning and engine counts
Metascan Multiscanning runs several anti-malware engines and combines their outputs. This broadens the available signals because vendors use different signatures, heuristics, machine-learning models, and update schedules.
Engine counts depend on context. OPSWAT’s community API documentation references reports from 30-plus engines, while the current commercial Threat Prevention API table lists 10, 15, or 23 engines for its Standard, Professional, and Enterprise tiers. A marketing statement about the community corpus should not be used as the guaranteed engine count for a production contract.
More engines do not turn detections into independent votes or a mathematical probability. One engine can produce a false positive; many engines can share blind spots or related intelligence. Investigators should inspect labels, file identity, signature and reputation information, static findings, sandbox evidence, and the consequences of a wrong decision.
The previous article claimed a fixed detection rate and sub-200-millisecond reputation path without a Top5soft test record. OPSWAT publishes efficacy and service-level claims for selected packages, but those vendor measurements are not independent site results and should be evaluated in the quoted configuration.
Deep CDR and sanitized files
Deep Content Disarm and Reconstruction treats supported content as potentially malicious. It parses the file, removes or changes active and disallowed components, and produces a reconstructed version intended for continued use. This prevention model can address document-borne risk without waiting for a malware signature.
Sanitization is not the same as proving that the source file was malicious or clean. It applies a transformation policy. Macros, embedded objects, scripts, hyperlinks, metadata, animations, forms, or unsupported structures can be removed or changed depending on format and configuration. That is a security benefit when those features are not needed and a business limitation when they are.
Teams should test representative documents before placing CDR in a production upload path. Validation should cover visual fidelity, accessibility, digital signatures, comments, formulas, links, forms, media, archive handling, output size, downstream parsing, and what happens when a format is unsupported or processing fails.
The current product page says Deep CDR supports 200 common file types in its upload-security use case. Support can vary by license and operation, so the API’s supported-file-type response and the contracted feature list are stronger implementation evidence than a general marketing count.
Sandbox, DLP, and other layers
MetaDefender Cloud can add dynamic sandbox analysis for supported files. The web documentation currently identifies a Windows 10 sandbox path and notes that licensed file-size limits can differ. Sandbox results can expose runtime behavior and indicators that static scanners do not see.
As with every sandbox, non-observation is not proof of safety. Malware can require a specific environment, delay execution, check for analysis artifacts, depend on user interaction, use encrypted configuration, or fetch a later stage only for selected targets. Analysts should verify that the expected code path actually ran.
Proactive DLP can identify sensitive data and, for supported workflows, block or redact it. File-type verification can catch extension and content mismatches. Archive extraction exposes nested content to the selected policy. Vulnerability and metadata checks can add software and provenance context.
Each layer has its own support matrix, cost, latency, and failure modes. A production policy should record which operations are required, which are optional, what result blocks a file, whether a sanitized output is allowed, and how timeouts, encrypted archives, oversized files, or unavailable engines are handled.
Community limits, APIs, and licensing
Community access is intended for demonstration, prototyping, and personal use under OPSWAT’s published restrictions. Commercial use requires an appropriate MetaDefender Cloud or MetaDefender Core license. Every API request uses an API key, and free keys have limited daily and per-minute capacity.
The release notes show why old quota tables age badly: OPSWAT has changed limits for anonymous and registered free users. The account dashboard, response headers, licensing page, and current documentation should be treated as authoritative at implementation time. This review does not repeat the previous article’s mutually inconsistent 5, 10, and 25 scans per day.
Current commercial tiers vary by daily or monthly volume, file-size limits, request throttling, archive depth, engine count, service level, organizations, authentication, and private controls. OPSWAT publishes tier structure but directs buyers to licensing or sales rather than maintaining a dependable universal per-file price.
A quote request should specify file volume and bursts, maximum size, archive depth, required engines and technologies, private mode, retention, region, DLP, sandbox, mTLS, identity provider, support, availability, and overage behavior. The old claims of ten-cent files and a fixed local-appliance price were not supported and have been removed.
Cloud versus customer-controlled deployment
MetaDefender Cloud is convenient for Internet-accessible integrations because OPSWAT operates the service and its engines. Even in private modes, the file is temporarily processed in OPSWAT’s cloud infrastructure.
Organizations that cannot send files to an external cloud can evaluate the broader MetaDefender product family for local, private-cloud, or air-gapped deployment. Those options can include multiscanning, Deep CDR, vulnerability assessment, and other modules, but capabilities depend on the selected product, engines, hardware, and license.
On-premises deployment changes rather than eliminates operational work. The customer must size throughput, deploy updates and engines, manage certificates and keys, monitor health, handle queues and failures, protect stored files and reports, and maintain high availability where required.
Do not present MetaDefender Cloud’s community website and an air-gapped MetaDefender Core appliance as interchangeable plans. They have different trust boundaries, administration, procurement, and evidence for compliance.
Strengths and limitations
Documented strengths
- Combines reputation, multiscanning, CDR, sandboxing, DLP, and file checks as policy layers.
- Supports hash lookup before file upload.
- Offers paid Private Scanning and stronger Private Processing controls.
- Makes sanitized output available for supported Deep CDR workflows.
- Provides REST APIs and commercial tiers for automated file-upload pipelines.
- Has customer-controlled MetaDefender deployment options for restricted environments.
- Documents licensing, confidentiality, rate limits, and release changes in detail.
Important limitations
- Standard submission results and executables can participate in sharing.
- Private Scanning removes the file but can retain scan results.
- Deep CDR can alter legitimate active content and document behavior.
- Engine counts and capabilities vary by community access and commercial tier.
- Sandbox non-detection is not a safety guarantee.
- Community APIs are not licensed for ordinary production use.
- Pricing requires configuration-specific licensing or sales information.
- On-premises options require substantial deployment and update operations.
Alternatives and complementary tools
The source-verified file scanner guide separates reputation, multiscanning, behavior analysis, and document sanitization.
VirusTotal is a strong first hash-reputation lookup with a broad shared corpus. Its licensed Private Scanning reports omit partner antivirus verdicts, unlike its standard community reports.
Hybrid Analysis provides detailed public Falcon Sandbox reports for authorized malware-research samples. Its public service should not receive confidential files.
ANY.RUN emphasizes interactive sandbox sessions, which can help trigger dialogs and staged behavior. Community sessions are public; private work requires an appropriate plan.
Filescan.io focuses on rapid static analysis and emulation within the OPSWAT ecosystem. Its workflow and disclosure model still need to be evaluated separately.
Dangerzone locally reconstructs supported documents as PDFs. It lacks multiscanner and malware-behavior reporting, but avoids cloud submission of the source document.
Verdict
MetaDefender Cloud is most compelling as a configurable file-processing platform, not as a one-off score page. Multiscanning, Deep CDR, DLP, sandboxing, and private controls can be combined around the risk of a particular upload workflow.
The correct mode matters. Standard submissions contribute results and, for executables, potentially samples to OPSWAT’s sharing ecosystem. Private Scanning removes the file after analysis but retains results. Private Processing restricts results further and removes them from ordinary hash lookup. Sensitive integrations should enforce the selected mode by policy rather than rely on each uploader to remember it.
Test CDR fidelity and failure handling with representative business documents, confirm the exact engine and feature entitlement, and monitor quotas and API release notes. For files that cannot leave the organization, evaluate a customer-controlled MetaDefender deployment instead of treating the cloud’s private flag as an air gap.
Best fit: managed file-upload pipelines needing several prevention and analysis layers through one API.
Poor fit: casual users expecting unlimited free scans, workflows that cannot tolerate document transformation, or air-gapped processing without a separately deployed local product.
Frequently asked questions
Is MetaDefender Cloud free?
Limited community access exists for personal use, demos, and prototyping. Production and commercial integrations require an appropriate license.
Are standard uploads private?
No. Standard submission results can be shared with the cybersecurity community, and executables can be shared with malware-exchange partners. Select a paid private mode when authorized sensitive processing is required.
What is the difference between Private Scanning and Private Processing?
Private Scanning removes the file after analysis and does not share it, while results remain in the cloud database. Private Processing also restricts results to the submitting owner and keeps them out of ordinary hash lookup.
Does Deep CDR guarantee a harmless document?
No security layer can provide an absolute guarantee. Deep CDR removes or changes supported active content according to policy, and the reconstructed output must still be validated for the intended workflow.
Can MetaDefender run in an air-gapped environment?
The wider MetaDefender product family includes customer-controlled and air-gapped deployment options. That requires a separate product selection, license, infrastructure, and operations plan from MetaDefender Cloud.
