How accurate is the AI transcription?

Whisper large-v3 achieves near-human accuracy for clear audio. It handles accents, technical terms, and natural conversation with minimal errors.

What languages are supported?

Over 90 languages including English, Spanish, French, German, Chinese, Japanese, Arabic, Hindi, and many more with automatic detection.

Can I export as SRT subtitles?

Yes. The transcription includes timestamps which are automatically formatted into SRT subtitle files for use in video editors.

Does it work with background noise?

Yes, Whisper handles moderate background noise well. For best results, use audio with clear speech and minimal overlapping voices.

Can I transcribe video files?

Extract the audio track from your video first, or upload the audio portion. We support common audio formats from video exports.

How much does transcription cost?

Each transcription costs 0.5 credits regardless of audio length. New accounts receive free credits to try the service.

AI-Powered Transcription

Transcribe Audio to Text in Seconds

Convert any audio file to accurate text with Whisper AI. Supports 90+ languages with timestamps and SRT subtitle export.

90+ Languages
~4s Processing
0.5 Credit

Operational
0.5 cr · Whisper v3

A failed job costs nothing — credits are returned automatically. We keep photo and video results for 30 days, and guest files are deleted after 14 days.

Credit packs from $1.49, one-time. Credits never expire and no subscription is required.

What is Speech to Text?

Speech to Text is an AI tool in Pixelift that runs in the browser with no installation. Convert any audio file to accurate text with Whisper AI. Supports 90+ languages with timestamps and SRT subtitle export. Engine: Whisper v3. Cost: 0.5 credits per result; credits are bought once in packs from $1.49, no subscription, and they never expire. Files are processed on EU servers and never used to train models.

How It Works

Step 1
Upload Audio
Drag and drop an audio file — MP3, WAV, FLAC, OGG, or M4A up to 50MB.
Step 2
AI Transcribes
Whisper large-v3 processes your audio with near-human accuracy and timestamps.
Step 3
Get Your Text
Copy the transcript, download as text, or export SRT subtitles for video editing.

Why Use AI Transcription?

90+ Languages

Whisper large-v3 recognizes over 90 languages with automatic detection. From English to Japanese, Arabic to Polish — just upload and go.

Lightning Fast

Most files process in 3-8 seconds regardless of length. No waiting around for results — get your transcript almost instantly.

SRT Export

Get word-level timestamps automatically formatted into SRT subtitle files. Perfect for video editing and captioning workflows.

Use Cases

Transcription for every workflow

Video Subtitles

Generate accurate SRT files for YouTube, TikTok, and social media videos. Improve accessibility and engagement.

Podcast Notes

Turn podcast episodes into searchable text transcripts. Great for show notes, blog posts, and SEO content.

Meeting Minutes

Record meetings and let AI create detailed transcripts. Never miss action items or important decisions again.

Content Repurposing

Transform audio content into written articles, social posts, and documentation. Maximize every piece of content.

Frequently Asked Questions

How accurate is the transcription?

Whisper large-v3 achieves near-human accuracy for clear audio. It handles accents, technical vocabulary, and natural conversation with minimal errors.

What audio formats are supported?

MP3, WAV, FLAC, OGG, and M4A files up to 50MB. Most common audio and podcast formats work right out of the box.

Can I get SRT subtitles?

Yes. Click the Export SRT button after transcription. Timestamps are formatted automatically for use in any video editor.

How many languages are supported?

Over 90 languages with automatic detection. Select a specific language for better accuracy or leave it on auto-detect.

How long does transcription take?

Most files complete in 3-8 seconds. The incredibly-fast-whisper model is optimized for speed without sacrificing accuracy.

More tools

Video & audio

See all

Ready to Transcribe?

Upload your audio and get an accurate transcript in seconds. No software to install — works right in your browser.

Transcribe Audio See pricing

Call it from code

One endpoint, key in the header. The same tool that runs on this page — no browser, no clicking.

curl -X POST https://pixelift.pl/api/v1/generations \
  -H "Authorization: Bearer pk_live_..." \
  -H "Idempotency-Key: $(uuidgen)" \
  -F "tool=speech-to-text" \
  -F "audio=@nagranie.mp3" \
  -F "language=polish"

Response: { "id": "cmk8…", "tool": "speech-to-text", "status": "succeeded", "assets": [{ "kind": "output", "url": "/uploads/…" }], "result": { "text": "…", "creditsUsed": 0.5 } }

A retry with the same Idempotency-Key header is never charged twice; status and the event stream live at /api/v1/generations/{id}. The tool's direct route still works: POST /api/ai-audio/speech-to-text

Request fields

`tool`	string	required	Narzedzie (v1)
`audio`	file	required	Nagranie
`language`	string	optional	Pelna nazwa jezyka po angielsku albo „None" dla autodetekcji