Language Support Tables¶
In a nutshell
The per-language STT/TTS/parser coverage tables for OVOS: which languages have a bundled offline speech model, and which have a number/date parser module. Use it to look up one language. For the narrative on how language support works and how to improve it, see Language Support.
Supported Languages¶
ovos-config autoconfigure reads the recommendation files in the recommends/ directory of
your installed ovos-config. These files are the authoritative list. The tables below show
what each release channel ships.
Stable channel¶
The stable channel installs ovos-config 1.2.2. It recommends offline speech recognition for
11 languages: ca-es, da-dk, de-de, en-us, es-es, eu-es, fr-fr, gl-es, it-it,
nl-nl and pt-pt. It has no --gpu tier.
| Offline STT plugin | Languages |
|---|---|
ovos-stt-plugin-citrinet |
ca-es, de-de, fr-fr, it-it, nl-nl |
ovos-stt-plugin-fasterwhisper |
da-dk, en-us, pt-pt |
ovos-stt-plugin-fasterwhisper-zuazo |
es-es, eu-es, gl-es |
Testing channel¶
The testing channel installs ovos-config 2.1.1. It covers the same 11 languages and adds a
--gpu tier.
| Tier | STT plugin | Languages |
|---|---|---|
| Offline | ovos-stt-plugin-citrinet |
ca-es, de-de, en-us, es-es, fr-fr, it-it, nl-nl, pt-pt |
| Offline | ovos-stt-plugin-fasterwhisper |
da-dk, eu-es, gl-es |
| GPU | ovos-stt-plugin-fasterwhisper |
every language except eu-es |
| GPU | ovos-stt-plugin-HiTZ |
eu-es |
Alpha channel¶
Alpha channel only
The alpha ovos-config recommends offline speech recognition for 69 languages and a GPU
tier for 60. Both tiers use ovos-stt-plugin-onnx-asr,
one plugin that runs every model architecture in the table.
int8 next to a model means
the recommendation uses a quantized CPU build. The GPU tier sets use_cuda: true.
Where the table shows offline TTS voices, the recommendation is
ovos-tts-plugin-phoonnx with one pinned
voice for each gender listed. A single-speaker Facebook MMS voice counts as male.
autoconfigure adds a voice only when its gender is known.
"none" means the release has no recommendation for that tier. autoconfigure skips that
part of the configuration and tells you so. Use an online STT plugin, or configure a
multilingual model by hand, for example
ovos-stt-plugin-server pointed at a
self-hosted STT server.
| Language | Offline STT model | GPU STT model | Offline TTS voices |
|---|---|---|---|
| arb | OpenVoiceOS/stt_ar_fastconformer_hybrid_large_pcd_v1.0_onnx |
none | female, male |
| arg | none | none | female, male |
| as-IN | OpenVoiceOS/ai4bharat-indicconformer-as-onnx |
OpenVoiceOS/ai4bharat-indicconformer-as-onnx |
male |
| ast | none | none | female, male |
| be-BY | OpenVoiceOS/nvidia-be-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-be-conformer-transducer-large-onnx |
none |
| bn-IN | OpenVoiceOS/ai4bharat-indicconformer-bn-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-bn-onnx |
male |
| brx-IN | OpenVoiceOS/ai4bharat-indicconformer-brx-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-brx-onnx |
none |
| ca-ba | none | none | female, male |
| ca-ES | OpenVoiceOS/stt-ca-es-conformer-transducer-large-onnx |
OpenVoiceOS/whisper-large-v3-ca-punctuated-3370h-onnx |
female, male |
| ca-nw | none | none | female, male |
| ca-va | none | none | female, male |
| da-DK | OpenVoiceOS/nvidia-parakeet-rnnt-110m-da-dk-onnx (int8) |
nemo-canary-1b-v2 |
male |
| de-DE | nemo-parakeet-tdt-0.6b-v3 (int8) |
nemo-canary-1b-v2 |
female, male |
| doi-IN | OpenVoiceOS/ai4bharat-indicconformer-doi-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-doi-onnx |
none |
| en-GB | none | none | female, male |
| en-US | nemo-parakeet-tdt-0.6b-v3 (int8) |
nemo-canary-1b-v2 |
female, male |
| eo | OpenVoiceOS/nvidia-eo-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-eo-conformer-transducer-large-onnx |
none |
| es-ES | OpenVoiceOS/parakeet-rnnt-1.1b-cv17-es-ep18-1270h-onnx (int8) |
OpenVoiceOS/parakeet-rnnt-1.1b-cv17-es-ep18-1270h-onnx |
female, male |
| et-EE | OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-et-onnx (int8) |
OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-et-onnx |
none |
| eus | OpenVoiceOS/stt-eu-conformer-transducer-large-onnx |
OpenVoiceOS/hitz-eu-conformer-transducer-large-v2-onnx |
female, male |
| fa-IR | OpenVoiceOS/nvidia-fa-fastconformer-hybrid-large-onnx (int8) |
OpenVoiceOS/nvidia-fa-fastconformer-hybrid-large-onnx |
female, male |
| fo-FO | OpenVoiceOS/carlosdanielhernandezmena-stt_fo_quartznet15x5_sp_ep163_100h_onnx |
none | none |
| fon-BJ | OpenVoiceOS/misterkissi-whisper-small-fongbe-onnx |
OpenVoiceOS/misterkissi-whisper-small-fongbe-onnx |
male |
| fr-FR | nemo-parakeet-tdt-0.6b-v3 (int8) |
nemo-canary-1b-v2 |
female, male |
| fy-NL | none | none | female, male |
| ga-IE | OpenVoiceOS/misterkissi-w2v2-lg-xls-r-300m-ga-onnx |
OpenVoiceOS/misterkissi-w2v2-lg-xls-r-300m-ga-onnx |
none |
| gl-ES | OpenVoiceOS/proxectonos-gl-conformer-ctc-large-onnx |
onnx-community/whisper-large-v3-turbo |
female, male |
| gu-IN | OpenVoiceOS/ai4bharat-indicconformer-gu-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-gu-onnx |
male |
| hi-IN | OpenVoiceOS/ai4bharat-indicconformer-hi-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-hi-onnx |
female, male |
| hr-HR | OpenVoiceOS/nvidia-hr-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-hr-conformer-transducer-large-onnx |
none |
| ht-HT | OpenVoiceOS/misterkissi-whisper-small-haitian-creole-onnx |
OpenVoiceOS/misterkissi-whisper-small-haitian-creole-onnx |
male |
| hy-AM | OpenVoiceOS/stt_hy_fastconformer_hybrid_large_pc_onnx |
none | none |
| is-IS | OpenVoiceOS/carlosdanielhernandezmena-stt_is_quartznet15x5_ft_ep56_875h_onnx |
none | none |
| it-IT | nemo-parakeet-tdt-0.6b-v3 (int8) |
nemo-canary-1b-v2 |
female, male |
| ja-JP | OpenVoiceOS/nvidia-parakeet-tdt_ctc-0.6b-ja-onnx (int8) |
OpenVoiceOS/nvidia-parakeet-tdt_ctc-0.6b-ja-onnx |
female, male |
| ka-GE | OpenVoiceOS/stt_ka_fastconformer_hybrid_large_pc_onnx |
none | none |
| kab | OpenVoiceOS/nvidia-kab-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-kab-conformer-transducer-large-onnx |
male |
| kk-KZ | OpenVoiceOS/stt_kk_ru_fastconformer_hybrid_large_onnx |
none | none |
| kn-IN | OpenVoiceOS/ai4bharat-indicconformer-kn-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-kn-onnx |
male |
| ko-KR | OpenVoiceOS/stt_kr_citrinet1024_PublicCallCenter_1000H_onnx |
none | none |
| kok-IN | OpenVoiceOS/ai4bharat-indicconformer-kok-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-kok-onnx |
none |
| kpo-TG | OpenVoiceOS/misterkissi-whisper-small-kposo-onnx |
OpenVoiceOS/misterkissi-whisper-small-kposo-onnx |
none |
| ks-IN | OpenVoiceOS/ai4bharat-indicconformer-ks-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-ks-onnx |
none |
| mai-IN | OpenVoiceOS/ai4bharat-indicconformer-mai-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-mai-onnx |
male |
| mg-MG | OpenVoiceOS/misterkissi-whisper-small-malagasy-onnx |
OpenVoiceOS/misterkissi-whisper-small-malagasy-onnx |
male |
| ml-IN | OpenVoiceOS/ai4bharat-indicconformer-ml-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-ml-onnx |
male |
| mni-IN | OpenVoiceOS/ai4bharat-indicconformer-mni-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-mni-onnx |
none |
| mr-IN | OpenVoiceOS/ai4bharat-indicconformer-mr-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-mr-onnx |
male |
| mt-MT | OpenVoiceOS/carlosdanielhernandezmena-stt_mt_quartznet15x5_sp_ep255_64h_onnx |
none | none |
| ne-NP | OpenVoiceOS/ai4bharat-indicconformer-ne-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-ne-onnx |
none |
| ngl-MZ | OpenVoiceOS/misterkissi-whisper-small-lomwe-onnx |
OpenVoiceOS/misterkissi-whisper-small-lomwe-onnx |
male |
| nl-NL | nemo-parakeet-tdt-0.6b-v3 (int8) |
nemo-canary-1b-v2 |
female, male |
| oci | none | none | female, male |
| or-IN | OpenVoiceOS/ai4bharat-indicconformer-or-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-or-onnx |
none |
| pa-IN | OpenVoiceOS/ai4bharat-indicconformer-pa-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-pa-onnx |
male |
| pl-PL | OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-pl-onnx |
OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-pl-onnx |
female, male |
| pt-BR | none | OpenVoiceOS/whisper-large-v3-pt-onnx |
female, male |
| pt-PT | nemo-parakeet-tdt-0.6b-v3 (int8) |
OpenVoiceOS/whisper-large-v3-pt-onnx |
female, male |
| ru-RU | OpenVoiceOS/nvidia-ru-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-ru-conformer-transducer-large-onnx |
female, male |
| rw-RW | OpenVoiceOS/nvidia-rw-conformer-transducer-large-onnx (int8) |
OpenVoiceOS/nvidia-rw-conformer-transducer-large-onnx |
male |
| sa-IN | OpenVoiceOS/ai4bharat-indicconformer-sa-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-sa-onnx |
none |
| sat-IN | OpenVoiceOS/ai4bharat-indicconformer-sat-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-sat-onnx |
none |
| sd-IN | OpenVoiceOS/ai4bharat-indicconformer-sd-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-sd-onnx |
none |
| sl-SI | OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-sl-onnx (int8) |
OpenVoiceOS/yuriyvnv-parakeet-tdt-0.6b-sl-onnx |
female |
| sn-ZW | OpenVoiceOS/misterkissi-whisper-small-shona-onnx |
OpenVoiceOS/misterkissi-whisper-small-shona-onnx |
male |
| st-ZA | OpenVoiceOS/misterkissi-whisper-small-sesotho-onnx |
OpenVoiceOS/misterkissi-whisper-small-sesotho-onnx |
none |
| ta-IN | OpenVoiceOS/ai4bharat-indicconformer-ta-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-ta-onnx |
male |
| te-IN | OpenVoiceOS/ai4bharat-indicconformer-te-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-te-onnx |
male |
| tig-ER | OpenVoiceOS/misterkissi-whisper-small-tigre-onnx |
OpenVoiceOS/misterkissi-whisper-small-tigre-onnx |
none |
| tl-PH | OpenVoiceOS/stt-tl-fastconformer-hybrid-large-onnx |
OpenVoiceOS/stt-tl-fastconformer-hybrid-large-onnx |
male |
| tn-ZA | OpenVoiceOS/misterkissi-whisper-small-setswana-onnx |
OpenVoiceOS/misterkissi-whisper-small-setswana-onnx |
none |
| uk-UA | OpenVoiceOS/stt_ua_fastconformer_hybrid_large_pc_onnx |
none | none |
| umb-AO | OpenVoiceOS/misterkissi-whisper-small-umbundu-onnx |
OpenVoiceOS/misterkissi-whisper-small-umbundu-onnx |
none |
| ur-PK | OpenVoiceOS/ai4bharat-indicconformer-ur-onnx (int8) |
OpenVoiceOS/ai4bharat-indicconformer-ur-onnx |
male |
| uz-UZ | OpenVoiceOS/asr-uz-fastconformer-large-onnx |
OpenVoiceOS/asr-uz-fastconformer-large-onnx |
none |
| vai-LR | OpenVoiceOS/misterkissi-whisper-small-vai-onnx |
OpenVoiceOS/misterkissi-whisper-small-vai-onnx |
none |
| vi-VN | OpenVoiceOS/nvidia-parakeet-ctc-0.6b-vietnamese-onnx (int8) |
OpenVoiceOS/nvidia-parakeet-ctc-0.6b-vietnamese-onnx |
male |
| zh-CN | OpenVoiceOS/stt_zh_citrinet_1024_gamma_0_25_onnx |
none | none |
A language that is in neither table has no bundled recommendation. Thai is one example:
ovos-config ships no th-* recommendation file, so autoconfigure -l th-th has nothing to
select. Configure Thai by hand until somebody contributes a recommendation.
Per-language status matrix¶
The STT and TTS columns summarize the alpha-channel table above. "yes" means a bundled offline recommendation exists for at least one voice or gender. "partial" means only some of it exists. "no" means the release has no bundled offline recommendation.
Number and date parsing come from the language modules actually
shipped in the installed ovos-number-parser / ovos-date-parser. A yes here means a
dedicated language module exists (numbers_<code>.py / dates_<code>.py), not that every
phrasing is covered.
This matrix tracks the original, longest-supported languages. See the Supported Languages table above for the full, larger set of languages with a bundled offline STT and/or TTS recommendation.
| Language | Offline STT | Offline TTS | Number parser | Date parser |
|---|---|---|---|---|
| en-US | yes | yes | yes | yes |
| en-GB | no | yes | yes | yes |
| en-AU | no | no | yes | yes |
| de-DE | yes | yes | yes | yes |
| fr-FR | yes | yes | yes | yes |
| es-ES | yes | yes | yes | yes |
| it-IT | yes | yes | yes | yes |
| nl-NL | yes | yes | yes | yes |
| pt-PT | yes | yes | yes | yes |
| pt-BR | no | yes | yes | yes |
| ca-ES | yes | yes | yes | yes |
| gl-ES | yes | yes | yes | yes |
| eu-ES | yes | yes | yes | yes |
| da-DK | yes | partial (male only) | yes | yes |
Both the number and date parsers cover far more languages than this original matrix, including several with no offline STT or TTS recommendation, for example Hebrew and Turkish. Adding a language module for parsing text is much cheaper than shipping a full speech model, so this column tends to run ahead of STT/TTS.
pt-BR has no bundled offline STT
Brazilian Portuguese has no bundled offline CPU speech-to-text recommendation.
ovos-config autoconfigure -l pt-br --offline skips STT. On the alpha channel,
--gpu recommends a Whisper model for pt-br. Without a GPU, use an online STT plugin,
or configure a multilingual offline model by hand.
A language can have speech support and no parser module. Japanese is one example. The alpha
ovos-config recommends offline STT and TTS for ja-jp, but ovos-number-parser and
ovos-date-parser have no Japanese module. Spoken numbers and dates in Japanese are not
parsed until somebody contributes one.
Read next: Language Support Related: STT Plugins · TTS Plugins · Language Selection (internals)