01 / On device
Compress before transmission.
- MCU
- ZNano encode
- Radio
ZNano · Embedded Lossless Compression
Compress where
every byte
matters.
Reduce structured telemetry before transmission and reconstruct the original payload exactly on decode.
Designed for IoT, smart metering and constrained environments where bandwidth, memory, CPU and energy matter.
View Technical Evidence ↗
AT THE DEVICE
01 / STRUCTURED PAYLOAD
7A
02
5F
91
08
3C
7A
02
5F
91
08
3D
7A
02
5F
92
08
3C
7A
02
5F
91
08
3E
ZNANO ENCODE
Compressed payload
TRANSMIT
DECODE
RECOVER
7A
02
5F
91
08
3C
7A
02
5F
91
08
3D
7A
02
5F
92
08
3C
7A
02
5F
91
08
3E
STRUCTURED TELEMETRY
PAYLOAD-LEVEL
EXACT LOSSLESS RECOVERY
DEVICE · GATEWAY · BACKEND
01 / The fleet effect
Small payloads become large systems.
A compact reading is only the beginning. Device count and transmission frequency multiply the data moving through your infrastructure.
ILLUSTRATIVE FLEET EXAMPLE
Payload bytes only
72
bytes / reading
×
1,000
devices
×
1,440
readings / day
×
365
days
Explore fleet size
Raw annual payload
37.84 GB
At an assumed 56.32% reduction
16.53 GB
Calculated scenario using the published mean as an assumption, not a forecast for your fleet. Decimal GB; transport overhead excluded. Actual reduction depends on the payload.
Reduce data at the source.
Before repetition multiplies the cost.
Fewer transmitted bytes can reduce airtime and network load, depending on the transport and deployment.
02 / From device to application
Compress the payload. Keep the network.
ZNano operates at payload level, allowing your existing transport and application infrastructure to remain in place.
- DEVICE Sensor / meter Structured reading
- ENCODE ZNano Compressed payload
- TRANSPORT Your network Radio, uplink or storage
- GATEWAY / BACKEND ZNano decode Original payload
Compatible payload contexts
Payload contexts, not certified integrations. Transport framing, security and application validation remain part of your system.
03 / Verified performance
Measured on structured telemetry.
The last completed public host rerun, dated 15 August 2026. The repository is refreshing its benchmark matrix; these figures remain scoped to the published verified rows.
23.46–90.85%
Verified structured reduction range
56.32%
Verified mean reduction
58.14%
Verified median reduction
STRUCTURED REDUCTION
0–100% scale
Mean 56.32%
0%
23.46% minimum · 90.85% maximum
100%
10 / 10
Structured rerun rows
SHA-256 verified
SHA-256 verified
4 / 4
Incompressible controls
SHA-256 verified
View Public Benchmarks ↗
SHA-256 verified
Rows with failed verification are excluded from these lossless headline metrics. Incompressible data can expand; a strict size limit needs a wrapper or bypass policy.
04 / Integration points
Device. Gateway. Backend.
Place the engine where it makes sense for your architecture.
02 / At the gateway
Compress, batch and forward.
- Field devices
- Gateway + ZNano
- Uplink
03 / In the backend
Decode, store, replay and verify.
- Network / storage
- ZNano decode
- Application
Native C / C++ SDK, static libraries, CLI workflows and thin wrapper services support different integration models.
05 / Repeated records
Built for repeated telemetry.
Concatenated multi-frame processing supports repeated structured records. Public evidence includes 6, 10, 20 and 40-frame cases, depending on the dataset.
Across time
FRAME 01
FRAME 02
FRAME 03
FRAME 04
FRAME 05
FRAME 06
One multi-frame stream
Records remain available for decode. Internal stream layout is not shown.
Selective frame access
Decode the reading you need.
Request a selected frame without reconstructing every earlier frame first.
SELECTED RECORD
FRAME 06
Retrieve · Replay · Inspect
Conceptual selector. The compressed stream must be buffered or in randomly readable storage. Selective-decode timing is not yet published.
06 / Embedded systems
Designed for constrained systems.
A small native engine, with documented SDK build targets across the Cortex-M family.
ARM CORTEX
M0
ARM CORTEX
M0+
ARM CORTEX
M3
ARM CORTEX
M4
ARM CORTEX
M4F
Published host rerun · 15 August 2026
- Mean encode
- 1.407 ms
- Mean decode
- 1.537 ms
- Maximum host RSS delta
- 64 KB
Darwin arm64 · 30 measured runs after 5 warmups. Host timing is not MCU timing; host RSS delta is not MCU RAM consumption.
Documented SDK builds
Choose the target.
Budget for the complete system.
Cortex-M0 / M0+, M3, M4 soft-float and M4F hard-float builds are documented. Profiled archive footprints and static harness measurements are available in the public MCU report.
Application buffers and integration memory must be budgeted separately.
Explore MCU build evidence ↗07 / Live payload evaluation
Watch a telemetry stream compress.
Select a device and run the existing stream simulation. Its 60-second animation replays the processed workload, then displays the final byte totals and SHA-256 verification from the server.
Loading stream devices...
Device
Sensor
Standard stream
0 B / 0 payloads
ZNano stream
0 B / 0 payloads
Receiver
Decode & verify
Reference stream
--
ZNano stream
--
Payload reduction
--
SHA-256 round trip
--
Record size
--
Compared / pending
--
View technical log
Custom payload test
Test your own payload.
Use the payload tester to validate ZNano behavior on custom hexadecimal payloads. The existing tester retains its access and verification workflow.
08 / Use cases
Built for machines that keep talking.
Smart metering
Electricity, water and gas readings.
Industrial monitoring
Temperature, pressure and machine telemetry.
Agriculture
Soil, irrigation and livestock sensors.
Fleet & logistics
Position, diagnostics and cold-chain monitoring.
Smart infrastructure
Traffic, parking and environmental readings.
Remote operations
Maritime, mining, offshore and isolated systems.
Suitability depends on payload structure, batching and integration. No battery-life or radio-range improvement is implied.
09 / Engineering evidence
Public evidence. Private implementation.
Explore the measured outcomes, integration model and engineering history behind ZNano.
Public benchmarks ↗
Verified rows, host timing and methodology.
Embedded integration ↗
Target builds and scoped footprint evidence.
Architecture & claims ↗
Supported capabilities and evidence boundaries.
Historical ZHex / Ripple-lineage benchmark
Proven first on real metering data.
An earlier generation was developed and integrated for a gas-meter telemetry proof of concept. That work evolved into the device-agnostic ZNano architecture.
Read Technology Lineage ↗
5 / 5
Valid historical round trips
63.48%
Average reduction on that historical set
Historical results are separate from current ZNano benchmarks. The compression core, protected SDK releases, internal tooling and customer integrations remain private.
Optional reversible payload scrambling
Scrambling is not encryption and does not provide confidentiality, integrity or authentication.
This optional payload transformation is reversible. It does not change radio modulation or provide packet-loss resilience.
Read the integration boundaries ↗10 / Small payloads. Large systems.
Reduce the bytes between device and application.
Lossless payload compression for embedded devices, gateways and telemetry pipelines, without requiring a new transport protocol.