✳ FREE · NO UPLOAD · SPANISH

Spanish audio to text, free and private

Drop a Spanish recording or video. The speech model runs on your device and gives you an editable transcript with timestamps, plus SRT and VTT subtitles.

Public-domain LibriVox recording. Load it, then press Transcribe to try the full tool.

Whisper base, the model this tool runs, scored a 9.9% word error rate on Spanish in the Whisper paper's FLEURS test (lower is better). Expect a clean draft; check names and numbers.

First use downloads the speech model and engine (about 97 MB, of which 72.5 MB is the model); later visits load it from your browser cache. Runs on your CPU; nothing is uploaded. Up to 100 MB and 2 hours per file; under 30 minutes is the comfortable range on a laptop.

To transcribe Spanish audio for free, drop the file above: Whisper base runs in your browser, so nothing is uploaded. Spanish is one of its strongest languages, with a 9.9% word error rate on the FLEURS test in the Whisper paper, so expect a clean draft that needs a proofread for names and rare words.

Checked by Koldflux · updated 2026-09-25 · tested on a public-domain Spanish recording

How accurate is Spanish transcription?

Word error rate from the Whisper paper (lower is better): FLEURS, read sentences recorded under the es_419 (Latin American Spanish) locale, and Common Voice 9, volunteers with varied microphones. This tool runs Whisper base. Koldflux Studio runs Whisper large-v3-turbo, which the paper does not measure; OpenAI reports large-v3 cuts errors by 10 to 20% against large-v2, and the turbo version trades a little of that accuracy for speed.

ModelFLEURS WERCommon Voice WERVerdict (FLEURS)
Whisper base (this tool)9.9%19.6%Good
Whisper small5.6%10.3%Good
Whisper medium3.6%6.9%Good
Whisper large-v23%5.6%Good
Sources: Whisper paper (Radford et al., 2022), Appendix D (checked 2026-09-25) · FLEURS dataset card (Google) (checked 2026-09-25) · Whisper large-v3 model card (checked 2026-09-25) · Whisper large-v3-turbo model card (checked 2026-09-25)

What we saw on a real Spanish recording

We ran the first 30 seconds of “La zorra y el espino” from Aesop’s fables (LibriVox, public domain) through the same files this page loads, with the language set to Spanish. One 30-second clip is an illustration, not a benchmark. Processing took 7.1 seconds on one CPU thread of our test server.

Las fábulas de Soppo, grabado para LibriVox.org por Karen Savage. Fábula número 9. La zorra y el espino. Una zorra saltaba sobre unos montículos, y estuvo de pronto a punto de caerse, y para evitar la caída, se agarró a un espino, pero sus púas le irieron las patas…

Spanish spelling and punctuation in the output

Whisper writes Spanish with accents and ñ, and usually adds commas and full stops. Numbers often come out as digits (“Fábula número 9”), which is what subtitles want but not always what a script wants.

Whisper has a single Spanish language setting: there is no separate option for Spain, Mexico or Argentina. The paper’s FLEURS number was measured on Latin American Spanish recordings; other accents are handled by the same model.

Spanish subtitles and translation

Value
Characters per line (Netflix)42 characters per line
Reading speed (Netflix)up to 17 characters per second (adult programs)
Tool default line length42 characters
Speech to English, Whisper base (BLEU, higher is better)12.4
Speech to English, Whisper large-v2 (BLEU)23.3

Tips for Spanish

  1. Spanglish: pick Spanish by handIf most of the talk is Spanish with English words mixed in, set the language to Spanish rather than auto-detect, so a stretch of English does not switch the whole window to English.
  2. Check silent letters and proper nounsThe errors that survive are the ones that sound identical: a missing h, b versus v, and names. Search the transcript for the names you know are in the recording.
  3. Subtitles for Spanish viewersNetflix’s Spanish guide allows 42 characters per line and about 17 characters per second for adults, slower than its English limit of 20. The subtitle checker flags anything above 20, so treat its warnings as a floor for Spanish.

Limits and when to use something else

Frequently asked questions

Is Spanish transcription really free here?

Yes. The model runs in your browser, so there is no server cost to pass on: no sign-up, no watermark and no minute limit beyond what your device can handle. Files up to 100 MB and 2 hours.

Does it work for Spain and Latin American Spanish?

Both use the same Spanish setting; Whisper has no regional variants. The paper measured 9.9% word errors on Latin American Spanish read speech (FLEURS) and 19.6% on Common Voice, which has more varied speakers and microphones.

Can it translate Spanish audio to English?

Yes: choose “Translate the speech to English”. With this small model the translation is a rough gist (BLEU 12.4 in the paper, against 23.3 for the largest model). For subtitles you publish, transcribe in Spanish and translate the text separately.

How long does a Spanish file take?

In our test one CPU thread processed 30 seconds of Spanish in 7.1 seconds. Your laptop may be faster or slower; with WebGPU the encoder runs on the graphics card.

FROM TRANSCRIPT TO POSTS

Turn one idea into a week of posts

Koldflux Studio (paid, from $19/month) transcribes Spanish uploads on our servers with Whisper large-v3-turbo, the large model family (the paper measures 3.0% for large-v2), then turns the recording into Shorts and carousel posts you review, approve and export. It does not publish or schedule for you.

  1. 1 · Recording, idea or script
  2. 2 · Pick the angles
  3. 3 · Review and approve
  4. 4 · Export Shorts and carousels
See Koldflux Studio
Koldflux Studio: choosing which angles (Opportunities) to turn into Shorts before anything is generated
Koldflux Studio: you choose the angles before anything is made.

Language pages

Each page has the tool preset to that language, measured accuracy, our own test on a real recording and subtitle rules.