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.

Distil-Whisper 3.5

Run speech recognition in your browser. Unlimited runs.

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

Transcribe English recordings locally in a browser with optional segment timestamps using a distilled Whisper model. Returns text and segment timing; it does not identify speakers.

License
MIT
Model released
Unknown
Source repository created
Dec 5, 2024
Model weights · Q4
725.9 MB

Best for

  • Draft English transcripts and subtitles for recordings.
  • Local English speech transcription with segment timing.

Keep in mind

  • English only; the multilingual tokenizer does not make this an evaluated multilingual model. Translation is not supported.
  • A 756-million-parameter model with large downloads and substantial memory needs; device performance varies.
  • Accents, noise, music, overlapping speakers and domain vocabulary can cause omissions, hallucinations or incorrect words.
  • Segment timestamps are approximate; word-level alignment and speaker labels are not offered.
  • Chunk boundaries can repeat or omit text. Silence can produce hallucinations.
  • Review transcripts before publication; do not rely on unreviewed output for consequential decisions.

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
distil-whisper/distil-large-v3.5
Release verification
Unknown
Release evidence
Exact original checkpoint public release date has not been verified. Repository creation, training, paper, artifact conversion and update dates are not release evidence.
Source updated
Apr 13, 2026
Browser model files
onnx-community/distil-large-v3.5-ONNX
Files updated
Jun 1, 2025
Artifact revision
d908cd7fba0fb226f14b14ad955f22797ecfe5d5

Input

Audio ≤500 MB · Video ≤2 GB

Sample result

Distil-Whisper Large v3.5

Settings

English transcription only; the inherited multilingual tokenizer is not evidence of multilingual model quality. English is the required selection for this English-distilled model.
Return segment timestamps or just text. Word alignment is not advertised for this ONNX export. Enable for segment timing used by subtitles; disable when only text is needed.
Advanced settings 3
Transcribe English audio. Translation is not supported. Transcription is required; this adapter does not offer translation.
Split long audio into chunks of 1–30 seconds. Shorter chunks reduce per-chunk work but lose context; longer chunks preserve context and use more memory.
Overlap per side; no more than half the chunk length. More overlap can help chunk boundaries but repeats computation; zero overlap is faster and may lose boundary words.
Record a clip, then run inference. This does not send live audio to the model.

Output

Your transcript starts here

Choose audio or try a sample, then run transcription. Copy the text or export subtitles when ready.

Benchmark evidence

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

Distil-Whisper Large v3.5

16 configurations tested
Passed2
Failed14
TestedNov 2, 2026
RuntimeTransformers.js 4.2.0

Quality criteria

{ "maximumWordErrorRate": 0.2, "rejectEmpty": true, "rejectGibberish": true, "requireValidSegmentTimestamps": true }

PrecisionDeviceResult
Q4GPU · WebGPUAcceleratedPassedFast
Q4CPU · WASMPassedVery slow
BNB4GPU · WebGPUFailed—
BNB4CPU · WASMFailed—
Q4F16GPU · WebGPUFailed—
Q4F16CPU · WASMFailed—
Q8GPU · WebGPUFailed—
Q8CPU · WASMFailed—
INT8GPU · WebGPUFailed—
INT8CPU · WASMFailed—
UINT8GPU · WebGPUFailed—
UINT8CPU · WASMFailed—
FP16GPU · WebGPUFailed—
FP16CPU · WASMFailed—
FP32GPU · WebGPUFailed—
FP32CPU · WASMFailed—
Test setup & sample details
Quality gate
{ "maximumWordErrorRate": 0.2, "rejectEmpty": true, "rejectGibberish": true, "requireValidSegmentTimestamps": true }
Tested
2026-11-02T14:56:28.712Z
Environment
{ "browser": "156.0.8078.79", "os": "Darwin", "release": "27.2.0", "arch": "arm64", "cpu": "Apple M3 Max", "runtime": { "name": "Transformers.js", "version": "4.2.0" }, "workerType": "module", "cache": "fresh disk-backed profile per case", "artifactTransport": "checksum-verified pinned local mirror", "wasmThreads": 1 }
Runtime
Transformers.js 4.2.0
Artifact revision
d908cd7fba0fb226f14b14ad955f22797ecfe5d5
Fixture
{
  "file": "model-samples/vosk-jfk.flac",
  "url": "https://assets.localai.tools/assets/benchmarks/vosk/63a4b1e4c1dc655ac70961ffbf518acd249df237e5a0152faae9a4a836949715/vosk-jfk.flac",
  "size": 1152693,
  "sha1": "537ddf75dff344f8d31c1cd6ebee23ee8362d921",
  "sha256": "63a4b1e4c1dc655ac70961ffbf518acd249df237e5a0152faae9a4a836949715",
  "durationSeconds": 11
}
Rerun when
model artifact revision change; runtime version change; fixture or quality-gate change; browser/backend behavior change
Per-case quality & failure details
[
  {
    "precision": "q4",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": true,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.11940298507462686,
          "referenceWords": 67,
          "outputWords": 67,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "good"
        }
      ]
    },
    "runtime": {
      "loadMs": 3959.2000000029802,
      "inferenceMs": 1309.6000000089407,
      "totalMs": 10187.29999999702,
      "grade": "excellent",
      "realTimeFactor": 0.11905454545535825
    }
  },
  {
    "precision": "q4",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": true,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.11940298507462686,
          "referenceWords": 67,
          "outputWords": 67,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "good"
        }
      ]
    },
    "runtime": {
      "loadMs": 7603.20000000298,
      "inferenceMs": 53786.59999999404,
      "totalMs": 179649.39999999106,
      "grade": "very-slow",
      "realTimeFactor": 4.889690909090367
    }
  },
  {
    "precision": "bnb4",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.208955223880597,
          "referenceWords": 67,
          "outputWords": 74,
          "longestRepeatedWordRun": 2,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 3141.2000000029802,
      "inferenceMs": 42814.59999999404,
      "totalMs": 136188.19999998808,
      "grade": "very-slow",
      "realTimeFactor": 3.892236363635822
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "bnb4",
    "device": "wasm",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.208955223880597,
          "referenceWords": 67,
          "outputWords": 74,
          "longestRepeatedWordRun": 2,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 4102,
      "inferenceMs": 51673.29999999702,
      "totalMs": 161451.5,
      "grade": "very-slow",
      "realTimeFactor": 4.697572727272457
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0.7727272727272727,
      "referenceWords": 22,
      "outputWords": 5,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "failed",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0.7727272727272727,
          "referenceWords": 22,
          "outputWords": 5,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.7611940298507462,
          "referenceWords": 67,
          "outputWords": 34,
          "longestRepeatedWordRun": 10,
          "gibberish": true,
          "validTimestamps": false,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 2436.0999999940395,
      "inferenceMs": 1904.5999999940395,
      "totalMs": 8336.29999999702,
      "grade": "excellent",
      "realTimeFactor": 0.17314545454491267
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "q4f16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 3986.3999999910593,
      "grade": "failed"
    },
    "failure": "Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: /mnt/vss/_work/1/s/onnxruntime/core/graph/graph_utils.cc:30 int onnxruntime::graph_utils::GetIndexFromName(const Node &, const std::string &, bool) itr != node_args.end() was false. Attempting to get index by a name which does not exist:InsertedPrecisionFreeCast_/layers.0/self_attn_layer_norm/Constant_output_0for node: /layers.0/self_attn_layer_norm/Mul/SimplifiedLayerNormFusion/\n"
  },
  {
    "precision": "q8",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.208955223880597,
          "referenceWords": 67,
          "outputWords": 74,
          "longestRepeatedWordRun": 2,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 3290.2999999970198,
      "inferenceMs": 46085,
      "totalMs": 142943.79999999702,
      "grade": "very-slow",
      "realTimeFactor": 4.1895454545454545
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "q8",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 4654.29999999702,
      "grade": "failed"
    },
    "failure": "Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: qdq_actions.cc:137 TransposeDQWeightsForMatMulNBits Missing required scale: model.decoder.embed_tokens.weight_merged_0_scale for node: model.decoder.embed_tokens.weight_transposed_DequantizeLinear"
  },
  {
    "precision": "int8",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.208955223880597,
          "referenceWords": 67,
          "outputWords": 74,
          "longestRepeatedWordRun": 2,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 4951,
      "inferenceMs": 44543.10000000894,
      "totalMs": 140080.5,
      "grade": "very-slow",
      "realTimeFactor": 4.04937272727354
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "int8",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 4709.899999991059,
      "grade": "failed"
    },
    "failure": "Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: qdq_actions.cc:137 TransposeDQWeightsForMatMulNBits Missing required scale: model.decoder.embed_tokens.weight_merged_0_scale for node: model.decoder.embed_tokens.weight_transposed_DequantizeLinear"
  },
  {
    "precision": "uint8",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.208955223880597,
          "referenceWords": 67,
          "outputWords": 74,
          "longestRepeatedWordRun": 2,
          "gibberish": false,
          "validTimestamps": true,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 3111.7999999970198,
      "inferenceMs": 43913,
      "totalMs": 134863.40000000596,
      "grade": "very-slow",
      "realTimeFactor": 3.992090909090909
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "uint8",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 4539.100000008941,
      "grade": "failed"
    },
    "failure": "Can't create a session. ERROR_CODE: 1, ERROR_MESSAGE: qdq_actions.cc:137 TransposeDQWeightsForMatMulNBits Missing required scale: model.decoder.embed_tokens.weight_merged_0_scale for node: model.decoder.embed_tokens.weight_transposed_DequantizeLinear"
  },
  {
    "precision": "fp16",
    "device": "webgpu",
    "status": "fail",
    "quality": {
      "wordErrorRate": 0,
      "referenceWords": 22,
      "outputWords": 22,
      "longestRepeatedWordRun": 1,
      "gibberish": false,
      "validTimestamps": true,
      "passed": false,
      "grade": "excellent",
      "cases": [
        {
          "fixtureId": "jfk",
          "wordErrorRate": 0,
          "referenceWords": 22,
          "outputWords": 22,
          "longestRepeatedWordRun": 1,
          "gibberish": false,
          "validTimestamps": true,
          "passed": true,
          "grade": "excellent"
        },
        {
          "fixtureId": "standard",
          "wordErrorRate": 0.7910447761194029,
          "referenceWords": 67,
          "outputWords": 29,
          "longestRepeatedWordRun": 5,
          "gibberish": true,
          "validTimestamps": false,
          "passed": false,
          "grade": "failed"
        }
      ]
    },
    "runtime": {
      "loadMs": 3865,
      "inferenceMs": 1006.2000000029802,
      "totalMs": 8840.70000000298,
      "grade": "excellent",
      "realTimeFactor": 0.0914727272729982
    },
    "failure": "Transcription quality/timestamp gate failed"
  },
  {
    "precision": "fp16",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 5079.5999999940395,
      "grade": "failed"
    },
    "failure": "Can't create a session. ERROR_CODE: 6, ERROR_MESSAGE: std::bad_alloc"
  },
  {
    "precision": "fp32",
    "device": "webgpu",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 3783.053,
      "grade": "failed"
    },
    "failure": "Failed to construct 'Response': The ArrayBuffer/ArrayBufferView size exceeds the supported range."
  },
  {
    "precision": "fp32",
    "device": "wasm",
    "status": "fail",
    "quality": "Not embedded",
    "runtime": {
      "loadMs": null,
      "inferenceMs": null,
      "totalMs": 3731.0329170000005,
      "grade": "failed"
    },
    "failure": "Failed to construct 'Response': The ArrayBuffer/ArrayBufferView size exceeds the supported range."
  }
]

Something not quite right?

Tell us what happened with Distil-Whisper 3.5. 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.