Sources and licenses

Palabra Lexi relies on open dictionaries and data. Here is what we use, where it shows up and under which license. We credit CC BY-SA material with a link to the source and the license, and share our own adaptations of it under the same terms.

• Wiktionary (via kaikki.org / wiktextract) — tap translations and the offline dictionary, noun gender and plural, IPA transcription, verb forms and conjugation charts, links in word webs. License: CC BY-SA 4.0. Wiktionary text is also available under the GFDL. Our offline dictionary, form tables and word webs are adaptations of Wiktionary (definitions shortened to one line, entries selected and cross-checked) and are shared under the same CC BY-SA 4.0 license.
• UniMorph — verb forms — cross-checking conjugation charts. License: CC BY-SA 3.0.
• Open English WordNet and Open Multilingual Wordnet — links in word webs. License: CC BY 4.0, CC BY 3.0, Apache-2.0. Open English WordNet — CC BY 4.0; Spanish — Multilingual Central Repository (MCR 3.0), CC BY 3.0; Greek wordnet — Apache-2.0.
• FrequencyWords (Hermit Dave, from OpenSubtitles subtitles) — word frequency and the order of the “1000 words” lists. License: CC BY-SA 4.0.
• Wikimedia Commons — native recordings under “pronunciation”. Each recording has its own license (usually CC BY-SA or CC BY): the author, license and a link to the file are shown next to its play button.
• Wikipedia, Vikidia, Wikivoyage, Wikisource, Wikinews, Wikibooks — excerpts in Topics (up to 1,200 characters, marked “abridged”). License: CC BY-SA 4.0, CC BY 2.5. Each excerpt links to its page and names its license: CC BY-SA 4.0 for the wiki projects, CC BY 2.5 for Wikinews; “public domain” only when the Wikisource page itself is marked so.
• “Double reading” texts — fables, stories and classic excerpts in the daily session. Public-domain works — the author and work are named under each text; translations, retellings and questions are ours.
• CMU Pronouncing Dictionary — checking English sound-alikes on our server (not shown on the site). License: BSD-2-Clause.
• Speech synthesis: Kokoro v1.0 and Piper — saying words aloud when the device has no voice for the language, and dialogues. License: Apache-2.0, MIT. Spanish and the dialogues — Kokoro v1.0 (Apache-2.0); English words — Piper en_US-ljspeech-medium (model MIT, LJ Speech data in the public domain); Greek is read only by the device's own voice.
• Fonts: Alegreya, Literata, IBM Plex Mono and others — site and app typography. License: OFL.

What is ours

The code, design, associations, scenes, examples, stories, songs, comics and the translations of reading texts are ours. Material from the sources above stays under its own licenses; our terms of use do not change that.

Spotted a wrong credit, or need one of our CC BY-SA data files? Write via the support form.

Last updated 7 October 2026.