UnpackDNA
Turn malicious files into clean, labeled datasets your models can actually train on.
UnpackDNA is the same content-analysis engine behind FileDNA, redirected from sanitization toward extraction: an intelligent file-processing pipeline that transforms batches of suspicious, malicious, or complex files into clean, structured, labeled datasets — ready for AI and machine-learning use, with a complete audit trail for every file it touches.
How It Works
UnpackDNA processes a batch through four stages, entirely in a static environment — nothing in the batch is ever executed.
1
True Format Detection
UnpackDNA looks at a file’s actual content, not its name or extension, to identify what it really is before anything else happens to it.
2
Deep Unpacking
Every suspicious element is extracted — files within files within files, at unlimited levels of embedding.
3
Content Discovery
Images, scripts, and macros are detected and decoded down to source, so nothing in the dataset stays opaque.
4
Secure Delivery
Everything is packaged into a clean, labeled dataset with a complete audit log of every step that produced it.
Key Capabilities
Built for teams that need real malicious-file data without the risk of running it.
Training-ready output
Clean, structured, labeled data with full processing history — usable directly for training language and detection models, not just a raw sample dump.
No live code execution
Every file is analyzed through proprietary, signature-free algorithms in a static environment. Nothing in the batch is ever executed to find out what it does.
Full accountability
Every action UnpackDNA takes on a file is logged, so the resulting dataset comes with a complete, auditable chain of custody.
You control your data
Deploy in the cloud, on-premises, air-gapped, or hybrid — the same pipeline, running wherever your data is allowed to live.
Who It’s For
ML and AI teams training threat-detection or content-classification models, who need large volumes of labeled malicious-file data without the risk of handling live samples.
Security researchers building content-intelligence datasets from real-world malicious and complex files, with full provenance for every sample.
High-security and air-gapped organizations that need the same unpacking pipeline running entirely inside their own network, or as a lightweight agent on mail gateways and file servers.
UnpackDNA shares FileDNA’s core content-analysis engine but serves a different job: FileDNA disarms a file so a person or an agent can safely use it. UnpackDNA extracts and labels it so a model can safely learn from it.