Dictado
AboutSign in

About Dictado

Content levels.

Dictado is built on openly-licensed recordings of native Latin American Spanish — no TTS, no scraped audio. Pick a level that matches your current comfort and the app handles the review scheduling automatically.

Intermediate

A2–B1

Longer sentences, still written for learners. Everyday topics, real grammar variety, slower than native speed.

Native

B2+

Authentic Spanish produced for native audiences. Real speakers, real speed, real register. Essential for ear training.

Native and intermediate are available today. Pick either (or both) in Settings.

Content reference

Every clip you hear comes from a source you can trace back to its original collection. When a license requires attribution, the player credits the speaker and shows the exact license on each clip.

Intermediate

1 source

  • Tatoeba (Spanish, short phrases)

    Per recording — CC0 / CC BY 4.0 / CC BY-SA 4.0 / CC BY-NC 4.0

    A filtered slice of the Tatoeba corpus. Phrases are capped at eight words and sixty characters, so the ear has room to breathe. Each recording carries its contributor’s license, shown in the player. Speaker usernames are displayed for CC BY / CC BY-NC clips because attribution is part of the license.

    Themes: Short, learner-oriented Spanish sentences — everyday chatter, simple questions and commands — contributed by volunteers. Not Mexico-specific; voices are drawn from contributors across the Spanish-speaking world.

    We admit CC0, CC BY 4.0, CC BY-SA 4.0, and CC BY-NC 4.0 recordings. CC BY-NC-ND is excluded. If Dictado ever offers a paid tier, the CC BY-NC subset would need to be removed or relicensed.

Native

3 sources

  • Google FLEURS (es_419)

    CC-BY 4.0 — attribution

    Short news sentences read cleanly by 1–3 different voices per phrase. Every sentence ships with a professional English translation, so you can check comprehension as you go. Register is formal (headlines, explainers) rather than conversational.

    Themes: News-register sentences from Wikinews, recorded in a studio by multiple native speakers and paired with human English translations. Mixed Latin American accents, not Mexican-only.

    Filtered to recordings 1.5–15 seconds long; sentences dominated by proper nouns (25%+ of content words) are removed.

  • Mozilla Common Voice (Mexican Spanish)

    CC0 — public domain

    Crowd-recorded volunteers across Mexico reading short sentences aloud. The voices are real and varied, the pacing is natural, and the licensing is the most permissive on the web. This is the largest slice of Dictado’s library.

    Themes: Read-aloud Wikipedia-style prose. About 84% falls into a ‘general’ bucket; restaurant / shop / greetings are thinner.

    Filtered to clips tagged with a Mexican accent, at least two up-votes, zero down-votes, and 1.5–6.0 seconds in length.

  • CIEMPIESS-LIGHT

    CC-BY-SA 4.0 — attribution, share-alike

    Excerpts from Mexican broadcasters — the register you’d hear on the street or on the radio, not in a textbook. Connected speech, dropped syllables, and regional slang show up here. Useful for advanced listening practice.

    Themes: Mexican radio and television broadcast speech. Faster, more reduced, and more colloquial than Common Voice.

    Share-alike license: any redistribution of this subset needs to carry the same license forward.