Galician public-domain literature (Rosalía, Pondal, Curros)
Stylistic context for Tier 3 only; never normative.
Credits and attributions
Faladoiro relies on open-licence Galician language resources created by universities, public research programmes and volunteers. We list them because their licences require attribution.
Responsible use of AI
Faladoiro is a Galician tutor with artificial intelligence. We are guided by fairness, reliability, privacy, inclusion, transparency and human accountability. It is practice, not CELGA certification and not a substitute for a human examiner. Corrections can be wrong: review what matters and tell us if you see an error.
Your data is not used to train any LLM
Faladoiro does not use what you write, speak, practise or your corrections to train any language model of ours, and we do not sell it as a corpus. Your account and history live on our server, for private use in providing the service. When a feature calls a provider (chat model or Azure for voice/pronunciation), only what that turn needs is sent, for inference, not to train their models, under the terms we require. Speech recognition and primary Galician TTS stay on the server.
Corpora and linguistic data
SLI Galician Corpora — CLUVI exports (ES–GL parallel)
Primary source for the interference lexicon and the administrative register track.
Lexicons and dictionaries
SLI Galnet — Galician WordNet 3.0
Tier 1 lexical-semantic authority.
hunspell-gl — Galician spellchecker
Tier 1 valid-form universe; server-side only (Partition B).
Models
Nos ASR wav2vec2 Galician model
Galician speech recognition.
Nos MT OpenNMT ES–GL
Candidate generation for Spanish interference.
Nos TTS Brais
Galician speech synthesis.
Tools and analysers
apertium-glg — Apertium linguistic data for Galician
Server-side linguistic annotation and lemma/form mapping.
Only sources marked in the shipping path are listed here. Pending or uncleared resources are intentionally excluded. When a source changes, regenerate the attribution register and this page.