nRF Edge AI Library changelog

See the list of changes for a specific release of the nRF Edge AI Library.

Release v3.0.0 (22 September 2026)

This release is tagged as NRF-EDGEAI-RELEASE-3.0.0 (internal release commit 1000c2a2793a409d9d90b4bfe635c459a0e4c90b). This is a major release.

The runtime major version was raised from 2 to 3, which makes it incompatible with solutions generated for the v2.x runtime.

This release contains the following breaking changes:
- Solutions and applications built for the 2.x runtime do not work with this release. Retrain your model and export a new solution with Nordic Edge AI Lab 3.0.0. nrf_edgeai_is_runtime_compatible() compares the major version of the runtime library against the solution, so a solution created for a 2.x runtime causes nrf_edgeai_init() to fail with NRF_EDGEAI_ERR_INCOMPATIBLE.
- Feature processing is now split into two separate pipeline stages: DSP feature extraction and feature scaling. The nrf_edgeai_interfaces_t runtime interface structure now has a new mandatory scale_features member of type nrf_edgeai_iface_scale_features_t, which every solution context must provide. Solutions exported with Nordic Edge AI Lab 3.0.0 populate this member automatically.
- The combined nrf_edgeai_process_features_<mode>_<input>_<output> interface family was removed and replaced by the following:
  - process_features, which now provides DSP extraction only, through nrf_edgeai_process_features_dsp_i8(), nrf_edgeai_process_features_dsp_i16(), nrf_edgeai_process_features_dsp_f32(), and nrf_edgeai_process_features_empty().
  - The new nrf_edgeai_scale_features_* family declared in nrf_edgeai_scale_features.h, which handles scaling.
  - Renamed scaling mode prefixes, where scale_vector_* is now input_vector_*, and scale_window_* is now input_window_*.
- The nrf_edgeai_t runtime context now has a state member, which changes the structure layout. Applications and exported solutions must be recompiled against the 3.0.0 headers.
Added:
- The public nrf_edgeai_process_features() API, allowing DSP feature extraction to be triggered independently of model inference.
- Runtime state tracking through nrf_edgeai_state_t, with composite readiness masks and the following state bit definitions:
  - NRF_EDGEAI_STATE_RT_INITIALIZED
  - NRF_EDGEAI_STATE_RT_INPUTS_COLLECTED
  - NRF_EDGEAI_STATE_RT_FEATURES_PROCESSED
  - NRF_EDGEAI_STATE_RT_INFERENCE_COMPLETED
  The runtime now validates its state on entry to the input, feature processing, and inference stages.
- Error codes for state violations: NRF_EDGEAI_ERR_UNINITIALIZED (-8) and NRF_EDGEAI_ERR_WRONG_STATE (-9).
- nrf_edgeai_dsp_features_ctx() to obtain read-only access to the DSP feature extraction context, so computed features can be used from the application.
Updated:
- Feature processing was split into two separate pipeline stages: DSP feature extraction and feature scaling.
- nrf_edgeai_run_inference() keeps the previous single-call workflow working. If features have not been processed yet, it invokes nrf_edgeai_process_features() internally and propagates its status code unchanged.
Fixed:
- nrf_edgeai_feed_inputs() now rejects calls on a context that was never successfully initialized, returning NRF_EDGEAI_ERR_UNINITIALIZED instead of forwarding the data to an uninitialized input window context.

Known issues 

There are no critical known issues identified for this release.

Compatibility 

Nordic Edge AI Lab solutions version: 3.0.0
Axon driver version: 2.0.0

Release v2.2.2 (03 July 2026)

This release is tagged as NRF-EDGEAI-RELEASE-2.2.2 (internal release commit d33025f8dcb96859626bdabc671b108983f1c0d8).

Added support for Axon driver version 1.3.0.
Fixed:
- Compatibility of the library with PicoLib in the nRF Connect SDK. _impure_ptr references are now removed from the audio microfrontend utils.

Known issues 

There are no critical known issues identified for this release.

Compatibility 

Nordic Edge AI Lab solutions version: 2.2.0 - 2.2.2
Axon driver version: 1.1.0 - 1.3.0

Release v2.2.1 (09 April 2026)

This release is tagged as NRF-EDGEAI-RELEASE-2.2.1 (internal release commit 2e7626e4b9d0bd84c5a07251da0d91927ab83cff).

Added support for Axon driver version 1.1.0.
Fixed:
- Argument handling in positive and negative sigma crossing rate feature extraction functions (nrf_edgeai_feature_pscr and nrf_edgeai_feature_nscr) across all supported integer and float types.
- Type casting for sigma_factor and lag parameters in time-domain DSP feature extraction functions to correctly use the declared nrf_dsp_sigma_factor_t and uint8_t types.
- Refactored DSP clipping functions for Q63-to-Q31 and Q63-to-Q15 fixed-point conversions to ensure correct saturation behavior.

Known issues 

There are no critical known issues identified for this release.

Compatibility 

Nordic Edge AI Lab solutions version: 2.2.0 - 2.2.1
Axon driver version: 1.1.0

Release v2.2.0 (06 March 2026)

This release is tagged as NRF-EDGEAI-RELEASE-2.2.0 (internal release commit 7a38f672a9e71949ba5f0ff43acbc7e865fb89f5).

Added:
- Axon model support across the nRF Edge AI library, including model type handling and runtime integration alongside existing Neuton support.
- Inference lifecycle APIs for initialization and deinitialization.
- Custom-domain feature extraction and improved audio mel processing and wake-word decoding flow.
Fixed Doxygen warnings and documentation issues.

Known issues 

There are no critical known issues identified for this release.

Compatibility 

Nordic Edge AI Lab solutions version: 2.2.0
Axon driver version: 0.7.0 - 1.0.1

Release v1.0.0 (05 December 2025)

This release is tagged as RELEASE-NRF-EDGEAI-1.0.0 (internal release commit 9501cebe17ec8d404298f8657fa3549c3ce5c453).

Added initial project release based on Neuton libc project.

Known issues 

There are no critical known issues identified for this release.

Compatibility 

Nordic Edge AI Lab solutions version: 1.0.0