LICENSES.md

# Third-party licences (DPLY-04)

Every third-party asset this project ships, fetches at runtime, or loads into a visitor's
browser, with the permission each one is used under and how that permission is honoured.

This document **merges** `docs/ASSETS.md` (the VRM, plan 01-01) and `docs/VOICEVOX-SETUP.md`
(the three VOICEVOX licence layers and the dictionary, plan 01-04); it does not re-research
either. Every runtime-library licence below was re-verified on 2026-09-06 against the package
registry and the upstream repository, and the VRM's row is verified automatically:
`tests/e2e/test_avatar_loop.py::test_vrm_meta_matches_licenses` reads the shipped file's own
embedded `VRMC_vrm.meta` on the deployed Space and compares it with the `vrm_meta` block in
this file, so the document and the binary cannot silently disagree.

The repository's own code is MIT (see the Space `README.md` front-matter). Nothing below is
covered by that licence.

## Summary

| Asset | Version | Author / rights holder | Source URL | Licence / terms URL | Permission relied on | Required credit string | How we comply |
|---|---|---|---|---|---|---|---|
| **VRM character** `avatar/assets/tutor.vrm` | `VRM1_Constraint_Twist_Sample` v1.0.1, VRM 1.0 | pixiv Inc. — `(c) 2022 pixiv Inc.` | https://github.com/vrm-c/vrm-specification/tree/master/samples/VRM1_Constraint_Twist_Sample | https://vrm.dev/licenses/1.0/ (VRM Public License 1.0, with per-file flags embedded in the file) | Embedded `VRMC_vrm.meta`: `allowRedistribution: true`, `avatarPermission: everyone`, `modification: allowModificationRedistribution`, `commercialUsage: corporation` | **none** (`creditNotation: unnecessary`); credited voluntarily as `VRM1_Constraint_Twist_Sample (c) 2022 pixiv Inc. - VRM Public License 1.0` | Committed through Git LFS unmodified; credit rendered in the footer, the About panel and the README; the one restriction (`allowAntisocialOrHateUsage: false`) is stated below; sourcing gate re-raised and closed in plan 01-10 (kept; `docs/ASSETS.md`) |
| **VOICEVOX CORE** (software) | `voicevox_core` 0.17.0 (Python wheel, abi3) | Hiroshiba Kazuyuki (ヒホ) / the VOICEVOX project | https://github.com/VOICEVOX/voicevox_core/releases/tag/0.17.0 | https://voicevox.hiroshiba.jp/term/ (VOICEVOX ソフトウェア利用規約); code and build artefacts MIT — https://github.com/VOICEVOX/voicevox_core/blob/0.17.0/LICENSE | Software terms: commercial and non-commercial use permitted; generated audio may be provided onward **subject to the flow-down obligation of clause 3**. MIT for the linked code | `VOICEVOX:<character>` — satisfied by the character row below; plus the terms notice wherever audio is obtainable | Installed from the official release URL pinned in `requirements.txt`; the wheel is **never vendored** (禁止事項 forbids unauthorised redistribution of the software); flow-down notice rendered next to the Replay control |
| **VOICEVOX ONNX Runtime** | `voicevox_onnxruntime` 1.23.2 | VOICEVOX project build of Microsoft ONNX Runtime | https://github.com/VOICEVOX/onnxruntime-builder/releases/tag/voicevox_onnxruntime-1.23.2 | MIT (both `VOICEVOX/onnxruntime-builder` and `microsoft/onnxruntime`) | MIT | none | Fetched lazily at runtime into the gitignored `voicevox_runtime/` from the official release; never committed or redistributed (`docs/VOICEVOX-SETUP.md` § ONNX Runtime acquisition) |
| **VOICEVOX voice model** `voicevox/model/zundamon.vvm` | `voicevox_vvm` 0.17.0, file `0.vvm` (SHA-256 `ecd35374d4182cd883cba5040376f7f888cc6ba248b1c2f4cea07cdb34bb1318`) | the VOICEVOX project (model file); voices inside it belong to their character licensors | https://github.com/VOICEVOX/voicevox_vvm/releases/tag/0.17.0 | VOICEVOX 音声モデル 利用規約 — https://github.com/VOICEVOX/voicevox_vvm/blob/main/README.md (shipped as `README.txt` / `TERMS.txt` with the release) | Clause 2: 「アプリケーションに組み込んで再配布することができます」 — embedded redistribution is explicitly permitted; clause 4 flow-down; a credit that shows VOICEVOX was used is required | `VOICEVOX:ずんだもん` (only ずんだもん, style 3, is used; the other voices packaged in `0.vvm` are neither used nor credited) | Committed through Git LFS under clause 2; credit and flow-down notice rendered; the model is used only through VOICEVOX CORE |
| **ずんだもん** (character voice) | — (voice library shipped inside `0.vvm`) | SSS LLC (合同会社SSS) | https://zunko.jp/ | https://zunko.jp/con_ongen_kiyaku.html (音源利用規約, one document covering nine characters) | Commercial and non-commercial use of generated audio permitted **with credit**; app use requires the credit on an introduction/about screen | **`VOICEVOX:ずんだもん`** — exact, ASCII colon, no spaces | Persistent footer credit + About panel on first paint (server-delivered, not injected); flow-down notice linking both terms URLs; no prohibited-use path exists (see below) |
| **Open JTalk dictionary** `voicevox/open_jtalk_dic_utf_8-1.11/` | `open_jtalk_dic_utf_8-1.11` | © 2009 **Nara Institute of Science and Technology** (NAIST), Japan | https://downloads.sourceforge.net/project/open-jtalk/Dictionary/open_jtalk_dic-1.11/open_jtalk_dic_utf_8-1.11.tar.gz | BSD-3-Clause — `voicevox/open_jtalk_dic_utf_8-1.11/COPYING` (reproduced below) | BSD-3-Clause redistribution in binary form, with the notice reproduced | none (attribution notice) | `COPYING` is committed alongside the dictionary files; the notice is reproduced in this file; the About panel names NAIST. Open JTalk itself (Nagoya Institute of Technology / HTS Working Group, modified BSD) is linked inside VOICEVOX CORE and is attributed too |
| **three.js** (incl. its `GLTFLoader` addon) | 0.185.1 | three.js authors (mrdoob et al.) | https://github.com/mrdoob/three.js | MIT — https://github.com/mrdoob/three.js/blob/master/LICENSE | MIT: redistribution permitted with the copyright and permission notice kept | none | **Vendored into this repository** (plan 01-10) as `avatar/vendor/three.mjs` and `avatar/vendor/GLTFLoader.mjs`, the package's `LICENSE` copied alongside as `avatar/vendor/LICENSE-three.txt`; served by the Space itself, no CDN in the render path. Provenance, hashes and the regeneration command: `avatar/vendor/README.md` |
| **@pixiv/three-vrm** | 3.5.5 | pixiv Inc. | https://github.com/pixiv/three-vrm | MIT — https://github.com/pixiv/three-vrm/blob/dev/LICENSE | MIT: redistribution permitted with the copyright and permission notice kept | none | **Vendored into this repository** (plan 01-10) as `avatar/vendor/three-vrm.mjs` (the build against three@0.185.1), the package's `LICENSE` copied alongside as `avatar/vendor/LICENSE-three-vrm.txt`; served by the Space itself |
| **@huggingface/transformers** (transformers.js) | 4.2.0 | Hugging Face | https://github.com/huggingface/transformers.js | Apache-2.0 — https://github.com/huggingface/transformers.js/blob/main/LICENSE | Apache-2.0 use in the visitor's browser | none | **Not vendored, deliberately** (plan 01-10): still loaded by the visitor's browser from `https://esm.sh/@huggingface/transformers@4.2.0` on the first accepted push-to-talk, because the library fetches its own ONNX Runtime WASM assets and the Whisper model from the Hugging Face CDN regardless, so vendoring the shim alone would remove no dependency (`avatar/vendor/README.md`). Nothing of it is redistributed by this repository, so no NOTICE reproduction is triggered |
| **Whisper ASR model** `onnx-community/whisper-base` (q4 ONNX) | `onnx-community/whisper-base`, converted from `openai/whisper-base` | OpenAI (model); onnx-community (ONNX conversion) | https://huggingface.co/onnx-community/whisper-base | Weights: Apache-2.0 per the `openai/whisper-base` model card (`license: apache-2.0`; the conversion card declares `base_model: openai/whisper-base` and no licence of its own). Source code of `openai/whisper`: MIT | Apache-2.0 use of the weights in the visitor's browser | none (attribution) | **Not in this repository**: the visitor's browser downloads it from huggingface.co on the first accepted push (135.8 MB); attributed here and in the About panel ("Whisper"). The "better accuracy" model `whisper-large-v3-turbo` is not offered yet |

Ten rows for the eight asset classes 01-VALIDATION.md names: the ONNX Runtime is listed
separately from VOICEVOX CORE because it is fetched by a different mechanism and under a
different licence, and the Whisper weights are listed because the visitor's browser downloads
them even though this repository never ships them.

## Machine-readable blocks

The deployed suite parses these. Keep the keys and the values exactly as the shipped assets
and the rendered page have them.

The VRM's embedded `VRMC_vrm.meta`, as read out of `avatar/assets/tutor.vrm` itself (VRM
1.0 key names; `thumbnailImage` is an image index and is omitted):

```vrm_meta
name: VRM1_Constraint_Twist_Sample
version: v1.0.1
authors: pixiv Inc.
copyrightInformation: (c) 2022 pixiv Inc.
licenseUrl: https://vrm.dev/licenses/1.0/
avatarPermission: everyone
commercialUsage: corporation
creditNotation: unnecessary
allowRedistribution: true
modification: allowModificationRedistribution
allowAntisocialOrHateUsage: false
allowExcessivelySexualUsage: true
allowExcessivelyViolentUsage: true
allowPoliticalOrReligiousUsage: true
```

The credit strings the page must render (`tests/e2e/test_avatar_loop.py::test_credits_visible`
checks the footer, the About panel and the server-delivered config for them):

```credits
voice: VOICEVOX:ずんだもん
avatar: VRM1_Constraint_Twist_Sample (c) 2022 pixiv Inc. - VRM Public License 1.0
```

## VRM character — `avatar/assets/tutor.vrm`

> **FINAL.** The sourcing gate armed in plan 01-01 was re-raised at plan 01-10's phase-exit
> checkpoint (2026-09-06) and the owner kept this file as the shipped character: the embedded
> meta permits redistribution, modification and avatar use with no credit owed, and the page
> credits pixiv Inc. regardless. Decision and rationale: `docs/ASSETS.md` § Sourcing decision.

VRM Public License 1.0 is a **template, not a fixed grant**: redistribution, avatar use,
modification and commercial use are per-file flags set by the licensor inside the model's own
`VRMC_vrm.meta`. Two files citing the same licence URL can grant materially different rights.
The facts above were therefore parsed out of this file's glTF JSON chunk (10,776,032 bytes,
SHA-256 `12c2b97e95e700783a6a550dc0eee2d7880aeedccef9ae67bc4c5a2f0f2631a2`), never off a README.

Permissions relied on, and what each one covers here:

| Flag | Value | What it licenses in this project |
|---|---|---|
| `allowRedistribution` | `true` | Committing the file to this public repository and serving it from the Space |
| `avatarPermission` | `everyone` | Any visitor uses it as the tutor avatar |
| `modification` | `allowModificationRedistribution` | The rest pose applied at mount, and any future retargeting, still redistributable |
| `commercialUsage` | `corporation` | The broadest tier; a public portfolio Space is covered unambiguously |
| `creditNotation` | `unnecessary` | No credit is required. One is rendered anyway |
| `allowAntisocialOrHateUsage` | **`false`** | The one restriction. A Japanese-language tutor has no such use path; stated so it is not silently dropped |

## VOICEVOX — three licence layers, read separately

The voice output involves three rights holders, and their terms differ in one place that looks
like a contradiction until both are read.

**1. Software** — VOICEVOX ソフトウェア利用規約, https://voicevox.hiroshiba.jp/term/. Commercial
and non-commercial use permitted. Clause 3 attaches a **flow-down obligation**: when audio
generated here is made available to others, those others must be bound to the same terms.
禁止事項 forbids redistributing the software in whole or in part without authorisation, so the
`voicevox_core` wheel and the ONNX Runtime binary are always **referenced** from their official
release URLs and never copied into this repository. Separately from the terms of use, the
VOICEVOX CORE **code and build artefacts** are MIT-licensed as of 0.17.0 (`LICENSE`, © 2021
Hiroshiba Kazuyuki; README § ライセンス). *Correction to earlier project notes:* prebuilt cores
below version 0.16 were under a different (LGPL v3 / commercial dual) licence, and that is the
version some of this project's planning text still describes; the shipped 0.17.0 is MIT.

**2. Voice model** — VOICEVOX 音声モデル 利用規約, in the `voicevox_vvm` README and shipped as
`README.txt` / `TERMS.txt` with the release. Clause 2 (quoted verbatim in the table above)
states that the model may be embedded in an application and redistributed — **embedded
redistribution is explicitly permitted**, which is why `voicevox/model/zundamon.vvm` is committed (through Git LFS) while the
software is not. That asymmetry is deliberate on VOICEVOX's side and on ours: the software terms
forbid redistributing the *software*; the model terms permit redistributing the *model* inside an
application. Clause 4 carries the same flow-down obligation, and a credit showing that VOICEVOX
was used is required.

**3. Character** — ずんだもん, SSS LLC, https://zunko.jp/con_ongen_kiyaku.html. Audio generated
with the ずんだもん voice library may be used commercially and non-commercially **provided the
credit `VOICEVOX:ずんだもん` is displayed**. Without the credit, per-character licensing is
**¥400,000 (+ tax) per character** — which is why the credit is not optional and why it is
rendered on first paint rather than behind a click. Placement (clause 2): for app use, on an
introduction/about screen, somewhere findable with a little looking
(「アプリの紹介画面などに記載をお願いします。（少し探せばわかる場所に）」); Phase 1 renders it in a
persistent footer beside the avatar **and** in the About panel, open on first load.

Prohibited uses under the character terms, recorded so later phases can check against them:
公序良俗に反する利用; 政治的・宗教的活動; 情報商材; **意図的な**虚偽情報の作成・拡散; 風俗営業;
反社会的勢力による利用. The falsehood clause is scoped to *intentional* creation or spread —
relevant once an LLM produces the avatar's lines (Phase 3): a tutor that is sometimes wrong is not
in scope of it; a product designed to mislead would be.

There is no escalation path and none should be sought: SSS LLC's 免責条項 2 states that they do
not, as a rule, answer "does my use qualify" questions
(「本ガイドラインに該当するかどうかのご質問については、原則お答えしておりません。」). The posture is:
read the guideline, comply visibly, document here.

The credit obligation survives an engine swap: the official VOICEVOX Q&A answers that using the
intermediate `AudioQuery` with another synthesiser still requires the credit
(「必要です。」). The viseme timeline in this project is built from the `AudioQuery`, so the credit
is owed on that basis alone.

**Pre-cleared swap-in: VOICEVOX Nemo.** If the character terms ever become inconvenient,
`n0.vvm` from the same `voicevox_vvm` 0.17.0 release (nine voices, style IDs 10000–10008) is a
two-line change. Credit is just `VOICEVOX Nemo`; the terms live entirely in the `voicevox_vvm`
README (https://voicevox.hiroshiba.jp/nemo/term/); there is no third-party rights holder and no
¥400,000 credit-omission clause. The cost is the loss of a recognisable character.

## Open JTalk dictionary — `voicevox/open_jtalk_dic_utf_8-1.11/`

BSD-3-Clause. The shipped `COPYING` reads, verbatim:

> Copyright (c) 2009, Nara Institute of Science and Technology, Japan.
> All rights reserved.
>
> Redistribution and use in source and binary forms, with or without modification, are permitted
> provided that the following conditions are met: Redistributions of source code must retain the
> above copyright notice, this list of conditions and the following disclaimer. Redistributions in
> binary form must reproduce the above copyright notice, this list of conditions and the following
> disclaimer in the documentation and/or other materials provided with the distribution. Neither
> the name of the Nara Institute of Science and Technology (NAIST) nor the names of its
> contributors may be used to endorse or promote products derived from this software without
> specific prior written permission.
>
> THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
> IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
> FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
> CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
> CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
> SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
> OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
> POSSIBILITY OF SUCH DAMAGE.

Two institutions are genuinely involved, and earlier research conflated them:

- The **dictionary data** committed here (`open_jtalk_dic_utf_8-1.11`, NAIST Japanese Dictionary /
  IPAdic lineage) is © 2009 **Nara Institute of Science and Technology** — the notice above is the
  one that must be reproduced for the files actually redistributed in this repository.
- **Open JTalk itself** — the engine whose dictionary format this is, linked inside
  `voicevox_core` — is by the **Nagoya Institute of Technology** and the **HTS Working Group**,
  under a modified BSD licence. It is not redistributed by this repository; it is attributed
  because the About panel names the engine.

## Runtime libraries loaded into the visitor's browser

| Library | Version | Licence | Copyright | Loaded from |
|---|---|---|---|---|
| three.js | 0.185.1 | MIT | © 2010-2026 three.js authors | `avatar/vendor/three.mjs` + `avatar/vendor/GLTFLoader.mjs`, served by the Space (`/gradio_api/file=avatar/vendor/…`); licence text `avatar/vendor/LICENSE-three.txt` |
| @pixiv/three-vrm | 3.5.5 | MIT | © 2019-2026 pixiv Inc. | `avatar/vendor/three-vrm.mjs`, served by the Space; licence text `avatar/vendor/LICENSE-three-vrm.txt` |
| @huggingface/transformers | 4.2.0 | Apache-2.0 | © Hugging Face | `https://esm.sh/@huggingface/transformers@4.2.0` — the one remaining CDN load, on the push-to-talk path only, deliberately (see the table above and `avatar/vendor/README.md`) |

The two rendering libraries were **vendored by plan 01-10**: `scripts/vendor_modules.py`
downloads the exact esm.sh builds the browser executed before (three@0.185.1's es2022 build, its
GLTFLoader, and @pixiv/three-vrm@3.5.5 built against that three), rewrites their imports so all
three share one `./three.mjs`, asserts no absolute or root-relative import survives, and copies
each package's `LICENSE` beside the modules — MIT permits exactly that redistribution provided
the notice is kept. `avatar/vendor/README.md` records every file's SHA256, byte size and source
URL; `tests/test_vendor.py` fails if a module is hand-edited or a CDN URL returns to the render
path. transformers.js stays on the CDN because its own runtime assets and the Whisper model come
from the Hugging Face CDN whatever the shim's origin; the render path is what must survive a CDN
outage, and it now does.

The ASR model (`onnx-community/whisper-base`, q4) is downloaded by the browser from
huggingface.co on the first accepted push-to-talk utterance and cached by the browser; it is
never served from this repository or the Space. `docs/ASR-TIERS.md` records why this model and
this quantisation.

## Not used, and why

- **`AvatarSample_A`, `AvatarSample_B` and `AvatarSample_C` are NOT CC0.** Copyright is retained
  and VRoid Hub's conditions apply; VRoid's own FAQ prohibits redistributing them under a CC0
  designation. The claim is repeated confidently across tutorials and blog posts, and it is
  false. Recorded here so it is not reintroduced into this repository.
- CC0 **cannot be set** on VRoid Hub at all, so any "CC0 on VRoid Hub" claim is false by
  construction; and enabling "Allow third-party usage" on VRoid Hub **overwrites** a model's
  embedded licence with the Hub's own conditions, so a Hub-sourced file's embedded flags may not
  say what its author intended. Both are reasons the VRM row above reads the flags out of the
  shipped file.
- **Hosted ASR on accelerated hardware (tier D)** is deliberately not built: it would burn the
  visitor's ZeroGPU quota, contradicting SC-4. `tests/test_transport_seam.py` fails if one appears.
- **Style-Bert-VITS2** (AGPL-3.0), **pykakasi** (GPL-3.0) and **edge-tts** (undocumented
  endpoint) were rejected at stack selection for licence reasons and do not appear anywhere in
  the tree. Python dependencies that are not assets (`gradio`, `spaces`, `huggingface_hub`,
  `numpy`) are ordinary Apache-2.0 / BSD packages installed from PyPI and are not redistributed.

## Completeness review (plan 01-10 Task 3, 2026-09-06)

The eight-asset checklist from `01-VALIDATION.md`, re-read against the sources in this tree by
the execution agent on the owner's instruction (the owner asked for the manual rows to be done on
their behalf; the legal judgement is therefore the agent's reading, recorded so the owner can
overrule it). **Result: no gaps.**

| # | Asset | Row present | Checked against |
|---|---|---|---|
| 1 | VRM character | yes | `docs/ASSETS.md` (same source, licence URL and embedded flags); `tests/e2e/test_avatar_loop.py::test_vrm_meta_matches_licenses` green on the deployed file. Sourcing gate closed: the file is final |
| 2 | VOICEVOX software | yes | `docs/VOICEVOX-SETUP.md` (terms URL `voicevox.hiroshiba.jp/term/`, MIT at tag 0.17.0); `requirements.txt` installs exactly the 0.17.0 release wheel from the official URL |
| 3 | VOICEVOX voice model (`.vvm`) | yes | `voicevox_vvm` 0.17.0 release in `docs/VOICEVOX-SETUP.md`; `sha256sum voicevox/model/zundamon.vvm` = `ecd35374…1318`, the value in the row; the file is an LFS object |
| 4 | ずんだもん character terms | yes | `docs/VOICEVOX-SETUP.md` (`zunko.jp/con_ongen_kiyaku.html`); the credit string in the `credits` block is what `test_credits_visible` asserted on the Space |
| 5 | Open JTalk dictionary | yes | `voicevox/open_jtalk_dic_utf_8-1.11/COPYING` first line `Copyright (c) 2009, Nara Institute of Science and Technology, Japan.`; the Nagoya Institute of Technology / HTS Working Group attribution for Open JTalk itself, as `docs/VOICEVOX-SETUP.md` records; the About panel names both institutions |
| 6 | three.js | yes | `avatar/vendor/README.md` (three@0.185.1 es2022 build, SHA256 recorded, `tests/test_vendor.py` re-checks); `avatar/vendor/LICENSE-three.txt` is the MIT notice © 2010-2026 three.js authors |
| 7 | @pixiv/three-vrm | yes | `avatar/vendor/README.md` (3.5.5 built against three@0.185.1); `avatar/vendor/LICENSE-three-vrm.txt` is the MIT notice © 2019-2026 pixiv Inc. |
| 8 | @huggingface/transformers | yes | `avatar/asr.js` loads exactly `https://esm.sh/@huggingface/transformers@4.2.0`, the version in the row; not redistributed, so no NOTICE reproduction is owed; the Whisper weights it downloads have their own row |

Every cell of the summary table is filled (no `TODO`, `TBD` or empty cell); the two machine-readable
blocks are the ones the deployed suite parsed green. Nothing in this document is a claim the agent
is unwilling to publish.

## Verification record

| Check | Result (2026-09-06) |
|---|---|
| VRM `VRMC_vrm.meta` parsed from the shipped file | matches the `vrm_meta` block above, key for key |
| `voicevox_core` 0.17.0 licence | README § ライセンス at tag 0.17.0: MIT; `LICENSE` © 2021 Hiroshiba Kazuyuki; GitHub API `MIT` |
| `voicevox_vvm` clause 2 | present verbatim in the `main` README (line 21) |
| `VOICEVOX/onnxruntime-builder`, `microsoft/onnxruntime` | GitHub API: MIT, MIT |
| `three@0.185.1`, `@pixiv/three-vrm@3.5.5`, `@huggingface/transformers@4.2.0` | npm registry: MIT, MIT, Apache-2.0 (all three are also the current `latest`) |
| Vendored `avatar/vendor/*.mjs` and the two `LICENSE-*.txt` (plan 01-10) | SHA256 and byte size of each recorded in `avatar/vendor/README.md` from the fetch; `tests/test_vendor.py::test_vendored_hashes_match_readme` re-checks them on every quick loop; the licence texts are the packages' own `LICENSE` files from the npm tarballs (unpkg), both MIT |
| `openai/whisper-base` model card | HF API `cardData.license = apache-2.0`; `openai/whisper` repo MIT |
| `onnx-community/whisper-base` card | no licence field; `base_model: openai/whisper-base` |
| Open JTalk `COPYING` | first line: `Copyright (c) 2009, Nara Institute of Science and Technology, Japan.` |
| Terms URLs answer | voicevox.hiroshiba.jp/term/ 200 · zunko.jp/con_ongen_kiyaku.html 200 · voicevox.hiroshiba.jp/nemo/term/ 200 · vrm.dev/licenses/1.0/ 200 |