High-density compression
Store and move less data without losing a single byte. ZCaps applies adaptive statistical compression to reduce structured and machine-generated data while preserving exact recovery.
ZCaps
Next-generation compression for modern data infrastructure.
Adaptive lossless compression combining high-density storage, selective random access and parallel processing in a single engine.
Modern systems continuously ingest, store, move and retrieve structured and machine-generated data. ZCaps is designed around that lifecycle — not only around making a file smaller.
Explore ZMDCLESS STORAGE. SAME DATA.LOSSLESS
1.00 GB150.05 MB
Compressed footprintExact recovery
HIGH DENSITYSELECTIVE ACCESSPARALLEL PROCESSINGBYTE-EXACT RECOVERY
ZCaps capabilities
Selective random access
Retrieve what you need without rebuilding the whole archive. Independently addressable compressed chunks let targeted ranges be decoded without decompressing the entire dataset.
Parallel by design
Turn modern multi-core infrastructure into compression throughput. Independent compression units can be processed concurrently across files and workloads.
Data-adaptive
The data chooses the model. ZCaps analyzes the input structure and adapts its statistical model automatically rather than relying on a format-specific compression path.
Same engine. Different data. Different model.
01 / Selective random access
Compression should not make data harder to use.
Traditional archives can force applications to decode far more data than they actually need. ZCaps uses independently addressable compressed chunks to target the relevant portion and decode only the required chunk or chunks.
Traditional compressed stream01
Requested byte range ↓
ZCaps02
Requested byte range ↓
Conceptual comparison; archive behavior varies by format and configuration.
Direct chunk addressing · Selective decode · Exact reconstruction
02 / Server-scale processing
Built to scale with the server.
Modern infrastructure does not run on one core. ZCaps can process independent files and compression units concurrently, turning available CPU resources into higher aggregate throughput.
WORKLOAD DISTRIBUTION1 → 4 PROCESSES
Incoming independent files
Core 1ZCapsFile process 1
Core 2ZCapsFile process 2
Core 3ZCapsFile process 3
Core 4ZCapsFile process 4
3.16×aggregate compression throughput
2.69× aggregate decode throughput
File processesCompressDecode
1 process35.79 MB/s41.03 MB/s
4 processes113.25 MB/s110.22 MB/s
4 parallel file processes vs 1 · measured batch workload
Independent-file batch parallelism on ZMDC; these figures do not measure four-thread acceleration of a single file.
03 / Data-adaptive
The model adapts to the data — not the extension.
ZCaps uses adaptive contextual prediction and statistical entropy coding. The model evolves with the input while encoder and decoder maintain equivalent deterministic state, allowing exact byte-for-byte reconstruction.
DATA INPUT
Variable fields
AAAAAA
ABABAB
123123Structured recordsVariable fields
ADAPTIVE MODEL
Evolving statistical state
COMPRESSED OUTPUT
Exact reconstruction
Format agnostic
No dedicated compressor for each application dataset.
Online adaptation
The compression model evolves with the stream.
Lossless recovery
Decoded bytes reconstruct the exact original input.
No mandatory external dictionary.
04 / Measured on modern data
1 GB of modern data. 150 MB stored.
ZCaps was tested on the frozen ZMDC-1G v1 corpus: API traffic, telemetry, observability, logs, database changes, transactions, collaboration data, structured binary records and a high-entropy control.
0250 MB
150.05 MBfrom 1 GB
7 / 9workload families led
On ZMDC-1G v1, ZCaps produced the smallest aggregate output of the tested configurations and led seven of nine workload families.
The corpus was frozen before external compression comparisons.
Results are specific to the documented corpus, configuration and hardware. Consult the linked technical evidence for published methodology and settings.
05 / From ingest to insight
One engine across the data lifecycle.
ZCaps is designed as an infrastructure compression layer rather than only a desktop archive utility.
- 01Ingest
- 02Compress
- 03Store
- 04Select
- 05Decode
- 06Process
API events · Telemetry · Logs · Observability · Database / CDC · Transactions · Scientific data · Enterprise records
Data pipelines
Compress operational data as it enters the system.
Storage engines
Reduce the physical footprint of structured datasets.
Observability
Store more logs, traces and metrics.
Private infrastructure
Deploy compression entirely inside controlled environments.
Scientific workloads
Reduce large structured datasets while retaining selective access.
Product integration
Embed ZCaps directly inside applications and platforms.
06 / ZCaps for genomics
Built for very large scientific datasets.
Tested across genomic and bioinformatics workloads including VCF, SAM, FASTQ, FASTA, GTF, GFF3, MAF and other scientific formats.
07 / Integration
Built to become infrastructure.
ZCaps can sit beneath applications, storage systems and private data platforms as a proprietary lossless compression layer.
Embed
Integrate the engine directly inside products.
Deploy privately
Run it inside infrastructure controlled by the organization.
Scale
Parallelize workloads across modern compute resources.
Verify
Maintain exact byte-for-byte recovery.
ZCaps is already used within Zetako’s own product stack, including ZNode compression workflows.
Build with ZCaps
Your data does not need another archive utility.
It needs a better compression layer.
Adaptive compression. Selective access. Parallel processing. Exact recovery.