One endpoint. Full audit. Know exactly which files to keep.

Upload your audio files and choose exactly which checks to run - comparison, quality, pace - all in one API call.

Start Audit

What you get

Four checks - pick any combination, 10 credits each per file

music_note

Speaker Consistency

Every file is compared to the rest of your batch. Any file that sounds like a different voice gets automatically flagged.

  • Compare every file against the batch
  • Score how closely each file matches
  • Adjust how strict the flagging is

1 credit/file

checklist

Audio Quality

Each file is checked for glitches, clipping, silence gaps, and other audio problems.

  • Noise and glitch detection
  • Clipping and audio issues
  • Quality score for every file

1 credit/file

bolt

Speaking Speed

Measures how fast each file is spoken and flags any that are noticeably faster or slower than the rest.

  • Speed measured for every file
  • Compared against the batch average
  • Too-fast or too-slow files flagged

1 credit/file

description

Script Accuracy

Detects repetitions, spoken tags, and script deviations. Scripts are optional — provide them for more accurate spoken tag detection and word-level accuracy scores.

  • Repetition detection
  • Spoken tag detection
  • Accuracy scores & outlier flagging
  • Full transcription of each file

1 credit/file

POST /audit

The workflow

From upload to report in minutes

1

Upload your audio files

Drag and drop files into the dashboard or send them directly via the REST API using multipart upload. Supports MP3, WAV, OGG, FLAC, M4A, AAC, and more.

2

Choose your checks

Pick which checks to run - voice comparison, audio quality, speaking speed, script accuracy - or run all four. At least one must be enabled.

3

Get your report

You get a per-file breakdown showing scores, flagged issues, and which files need regenerating. Every response includes a link to your full report.

4

Regenerate only the bad ones

Take the list of flagged files and re-run them through your TTS pipeline. The rest of your batch is good to go - no wasted regeneration, no missed anomalies.

Quick start

One API call to analyze your batch

# Audit tracks - specify each check as true/false via query params
curl -X POST "https://api.ttsaudit.com/audit?comparison=true&quality=true&pace=false&scriptAccuracy=false" \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "files=@track1.mp3" \
  -F "files=@track2.mp3" \
  -F "files=@track3.mp3" \
  -F "accuracy=standard" \
  -F "deviationThreshold=0.15"

# Check balance
curl -s https://api.ttsaudit.com/balance \
  -H "X-API-Key: YOUR_API_KEY"

Sample response

Every API response includes a direct link to view the full report in your dashboard

Sample Response
{
  "score": 87.2,
  "summary": "2 of 4 checks passed. 1 file flagged for regeneration.",
  "fileCount": 3,
  "tracksToRegenerate": [
    { "file": "track-047.mp3", "reasons": [{ "check": "comparison", "message": "deviation 35.00%", "deviation": 0.35 }] }
  ],
  "checks": {
    "comparison": {
      "score": 82.7,
      "summary": "1 of 3 files flagged - 82% average similarity.",
      "similarityMatrix": [[1.0, 0.94, 0.61], [0.94, 1.0, 0.63], [0.61, 0.63, 1.0]],
      "tracks": [
        { "file": "track-001.mp3", "similarity": 0.94, "deviation": -0.02, "flagged": false },
        { "file": "track-002.mp3", "similarity": 0.93, "deviation": -0.03, "flagged": false },
        { "file": "track-047.mp3", "similarity": 0.61, "deviation": 0.35, "flagged": true }
      ]
    },
    "quality": {
      "score": 92.1,
      "summary": "Good audio quality - average score 92.1 across 3 files.",
      "tracks": [{ "score": 95.2, "flagged": false, "snrDb": 45.2, "issueCount": 1, ... }]
    },
    "pace": {
      "score": 87.5,
      "summary": "Consistent speaking speed - median 142.3 WPM, no outliers.",
      "unit": "wordsPerMinute",
      "tracks": [{ "wpm": 142.3, "wordCount": 284, "deviation": 0.02, "flagged": false }, ...],
      "median": 142.3
    },
    "scriptAccuracy": {
      "score": 94.2,
      "summary": "No tags spoken aloud. Average accuracy 94.2% across 3 files.",
      "spokenTagCount": 0,
      "repetitionCount": 0,
      "tracks": [
        { "accuracy": 96.0, "flagged": false, "tags": { "found": 2, "spoken": [] }, "repetitions": [] },
        { "accuracy": 92.0, "flagged": false, "tags": { "found": 0, "spoken": [] }, "repetitions": [] },
        ...
      ]
    }
  },
  "creditsUsed": 12,
  "timing": { "decode": 7.7, "checks": 5.1, "total": 13.2 }
}

Ready to find the bad files?

100 free credits on signup. No credit card required.

Start Audit API Docs