A model is still running

A model is still running or downloading. Leaving will stop the current operation. Stay on this page to let it finish.

Skip to content
LocalAI.toolsBeta

Join the waitlist

Get an email when the new apps and model runs launch.

No account needed.

NemotronLabs RE-USE Desktop

Restore speech from noisy or degraded recordings locally with NVIDIA NemotronLabs RE-USE. The WebAI worker returns a complete mono WAV.

Browser inferencePowered by WebAI · Private · Unlimited runs
About this model

Restore speech from noisy or degraded recordings locally with NVIDIA NemotronLabs RE-USE. The WebAI worker returns a complete mono WAV.

Created by
NVIDIA
Model released
Oct 22, 2026
Source repository created
Jul 22, 2026
Model weights · FP32
38.6 MB

Best for

  • Improve speech recordings before listening or transcription.

Keep in mind

  • Accepts normalized mono or stereo PCM or Float32 WAV up to 60 seconds at the documented input rates. Stereo is downmixed.
  • Requires WebGPU with at least 512 MiB per buffer and storage binding. Some convolutions and waveform processing run on the CPU.
  • Processes a complete recording with overlapping windows. This adapter returns audio after blending and does not stream intermediate results.
  • May suppress overlapping speakers and quiet speech or introduce artifacts. Inspect important recordings.
  • Laughter retention and real-world clipping, reverberation and codec recovery are not established for this browser configuration.

Processed on your deviceInference runs in this browser. Input/output sharing requires consent for each run. Basic run logs include device information and approximate connection location. Model and runtime files download from external hosts. Saved downloads may be cleared by your browser.

Sources & technical details
Original model
https://huggingface.co/nvidia/NVIDIA-NemotronLabs-RE-USE
Release verification
Verified
Release evidence
The pinned NVIDIA model card dates public release to October 22, 2026.
Source updated
Oct 22, 2026
Browser model files
https://huggingface.co/nvidia/NVIDIA-NemotronLabs-RE-USE
Files updated
Oct 22, 2026
Artifact revision
31f82f2efad1095034492f236e8dac9239ba713b

Input

AUDIO

Sample result

NemotronLabs RE-USE Desktop
Record a clip, then run inference. This does not send live audio to the model.

Output

Your enhanced audio will appear here

Choose a noisy speech recording or try the sample to reduce background noise.

Benchmark evidence

Saved worker evidence is separate from benchmarks run on your device. It does not guarantee speed or quality on every browser.

NemotronLabs RE-USE Desktop

8 configurations tested
Passed8
Failed0
TestedNov 15, 2026
Runtimereuse-webgpu 1df8e4beec9104b877eff102103287d3efed0e1e

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResult
FP32GPU · WebGPUAcceleratedPassedVery fast
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedVery slow
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-15T11:08:10.607Z
Environment
Chrome module worker, Apple M3 Max WebGPU, CPU waveform processing
Runtime
reuse-webgpu 1df8e4beec9104b877eff102103287d3efed0e1e
Artifact revision
31f82f2efad1095034492f236e8dac9239ba713b
Fixture
{
  "file": "model-samples/desktop/deepfilternet/white-input.wav",
  "sha256": "e77a0ca810068b4b4e7b17e5f928b028316659fdd1d4b67e16b502a0be0ae304",
  "sha1": "4bd9d3c1429dacfed089481eb336fafddf68b130",
  "size": 491600,
  "url": "https://assets.localai.tools/samples/v1/files/e77a0ca810068b4b4e7b17e5f928b028316659fdd1d4b67e16b502a0be0ae304/white-input.wav"
}
Rerun when
artifact revision; runtime source; fixture or quality gate; browser/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 15.991195189984984,
      "correlation": 0.9875815517745864
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 4269.300000011921,
      "totalMs": 11500.300000011921,
      "realTimeFactor": 1.6676953125046565,
      "grade": "offline",
      "performanceScore": 15.991195189984984,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 13.956277744678934,
      "correlation": 0.9836759397014314
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 3969.399999976158,
      "totalMs": 11200.399999976158,
      "realTimeFactor": 1.6005645161194186,
      "grade": "offline",
      "performanceScore": 13.956277744678934,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 19.99790221492827,
      "correlation": 0.9950373063524411
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 1865.699999988079,
      "totalMs": 9096.699999988079,
      "realTimeFactor": 1.013967391297869,
      "grade": "offline",
      "performanceScore": 19.99790221492827,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "correlation": 0.9935878092749513
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 4048.600000023842,
      "totalMs": 11279.600000023842,
      "realTimeFactor": 1.5814843750093133,
      "grade": "offline",
      "performanceScore": 0.9935878092749513,
      "performanceMetric": "clean waveform correlation",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "peak": 0
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 3.800000011920929,
      "totalMs": 7234.800000011921,
      "realTimeFactor": 0.003800000011920929,
      "grade": "real-time",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {},
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 169,
      "totalMs": 7400,
      "realTimeFactor": 11.572039942938659,
      "grade": "offline",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {},
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 4140,
      "totalMs": 11371,
      "realTimeFactor": 1.6171875,
      "grade": "offline",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 15.991195189984984,
      "correlation": 0.9875815517745864
    },
    "runtime": {
      "loadMs": 7231,
      "inferenceMs": 4030.900000035763,
      "totalMs": 11261.900000035763,
      "realTimeFactor": 1.5745703125139698,
      "grade": "offline",
      "performanceScore": 15.991195189984984,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "31f82f2efad1095034492f236e8dac9239ba713b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  }
]

NemotronLabs RE-USE Laughter Desktop

8 configurations tested
Passed8
Failed0
TestedNov 15, 2026
Runtimereuse-webgpu 1df8e4beec9104b877eff102103287d3efed0e1e

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResult
FP32GPU · WebGPUAcceleratedPassedVery fast
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedSlow
FP32GPU · WebGPUAcceleratedPassedVery slow
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-15T11:08:10.607Z
Environment
Chrome module worker, Apple M3 Max WebGPU, CPU waveform processing
Runtime
reuse-webgpu 1df8e4beec9104b877eff102103287d3efed0e1e
Artifact revision
1b1521dd659608e1e063ada83e0acfe1da473a8b
Fixture
{
  "file": "model-samples/desktop/deepfilternet/white-input.wav",
  "sha256": "e77a0ca810068b4b4e7b17e5f928b028316659fdd1d4b67e16b502a0be0ae304",
  "sha1": "4bd9d3c1429dacfed089481eb336fafddf68b130",
  "size": 491600,
  "url": "https://assets.localai.tools/samples/v1/files/e77a0ca810068b4b4e7b17e5f928b028316659fdd1d4b67e16b502a0be0ae304/white-input.wav"
}
Rerun when
artifact revision; runtime source; fixture or quality gate; browser/backend behavior
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 16.134779876105295,
      "correlation": 0.9879781595028334
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 4623.199999988079,
      "totalMs": 7895.300000011921,
      "realTimeFactor": 1.8059374999953435,
      "grade": "offline",
      "performanceScore": 16.134779876105295,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 13.913633055329537,
      "correlation": 0.9835188196800827
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 4041.399999976158,
      "totalMs": 7313.5,
      "realTimeFactor": 1.6295967741839348,
      "grade": "offline",
      "performanceScore": 13.913633055329537,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 19.71969741358448,
      "correlation": 0.9947115976366523
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 1960.800000011921,
      "totalMs": 5232.900000035763,
      "realTimeFactor": 1.0656521739195222,
      "grade": "offline",
      "performanceScore": 19.71969741358448,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "correlation": 0.9918855408368493
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 4152.5,
      "totalMs": 7424.600000023842,
      "realTimeFactor": 1.6220703125,
      "grade": "offline",
      "performanceScore": 0.9918855408368493,
      "performanceMetric": "clean waveform correlation",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "peak": 0
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 3.800000011920929,
      "totalMs": 3275.900000035763,
      "realTimeFactor": 0.003800000011920929,
      "grade": "real-time",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {},
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 154.10000002384186,
      "totalMs": 3426.2000000476837,
      "realTimeFactor": 10.55178316853696,
      "grade": "offline",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {},
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 4124.599999964237,
      "totalMs": 7396.699999988079,
      "realTimeFactor": 1.6111718749860302,
      "grade": "offline",
      "performanceScore": 1,
      "performanceMetric": "WAV integrity check",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "siSdrImprovementDb": 16.134779876105295,
      "correlation": 0.9879781595028334
    },
    "runtime": {
      "loadMs": 3272.100000023842,
      "inferenceMs": 3932.5,
      "totalMs": 7204.600000023842,
      "realTimeFactor": 1.5361328125,
      "grade": "offline",
      "performanceScore": 16.134779876105295,
      "performanceMetric": "SI-SDR improvement (dB)",
      "engine": "reuse-webgpu",
      "version": "1df8e4beec9104b877eff102103287d3efed0e1e",
      "artifactRevision": "1b1521dd659608e1e063ada83e0acfe1da473a8b",
      "requestedDevice": "webgpu",
      "device": "webgpu",
      "precision": "fp32",
      "fallback": false,
      "acceleration": "hybrid-webgpu-cpu-dsp"
    }
  }
]

Something not quite right?

Tell us what happened with NemotronLabs RE-USE. Loading, results, or the page itself—we’re here for all of it.

Attachments Optional

Up to 3 files · 50 MB each

Browser, OS, CPU/GPU, memory, screen, storage, and network details are included automatically. Model inputs, results, and browsing history are not collected.