Palash Debnath
43f1d46fe6
* fix(indextts): accept the config name upstream ships, and keep long text alive Two independent defects, both reported on a working IndexTTS 2.5 install. Install always failed. IndexTeam/IndexTTS-2.5 ships the model config as config.yaml — at the pinned revision d0aa86e7 and at HEAD; config_v2_5.yaml exists in no upstream revision. VoiceStudio demanded that name, so _weights_floor_ok never found it and the install died claiming 'the download was likely interrupted' when the download had been perfect. The only way through was to hand-rename the file. Both names are accepted now, in the installer and on the load path, so installs created with the workaround keep working without a reinstall. Long text was killed at 60s. infer() is one blocking upstream call that puts nothing on the wire, and IndexTTS was the only sidecar still on the 60s recv_timeout_s class default while pockettts and omnivoice-subprocess had both raised theirs. Raising the default alone does not fix it — which is why the reporter's RECV_TIMEOUT_S=3600 edit didn't help: progress frames are also what report activity to the GPU pool's execution clock (#1367), so a silent sidecar still trips the outer generate budget. The sidecar now heartbeats every 5s while infer() runs (and during the cold model construction), _send takes a lock so the beat thread can't interleave framing, and the deadline rises to 900s via OMNIVOICE_INDEXTTS_RECV_TIMEOUT_S. test_indextts25_health_requires_25_config_name asserted the bug — that a checkout holding only config.yaml is unhealthy — so it is rewritten to the corrected contract, including that a genuinely truncated download is still caught. Fixes #1611 * test(indextts): follow the installed config name in the sidecar loader tests Two more tests encoded the config_v2_5.yaml assumption, both asserting cfg_path against a directory where no config existed at all — so they were pinning the literal name rather than the resolution. They now lay down a real checkpoints/ tree and assert the resolved path, including that a checkout carrying the pre-fix hand-renamed config still resolves. Caught by the full suite; the targeted runs during development did not reach tests/backend/services/. * test(indextts): event-driven heartbeat tests, real interleaving proof, precedence pin Review round on #1619 — all four findings taken. - The docs line naming 0.5.1 is version-neutral now ('Earlier installs') — version labels are the owner's call. - The heartbeat tests waited on wall-clock sleeps; they now block on a per-write Event with a bounded deadline, so scheduler load can't flake them. - The _send test asserted the lock EXISTS — a tautology. It now drives four concurrent writers through a stream that yields between every byte and asserts every frame decodes; verified fail-before by removing the lock (torn frame) and pass-after. - The precedence test deleted config.yaml before creating the renamed one, so reversed precedence still passed. Both files now coexist for the assertion; verified fail-before by reversing _CFG_NAMES.
5.2 KiB
VoiceStudio — IndexTTS 2.5
IndexTTS 2.5 is an optional, multilingual voice-cloning engine for dubbing and expressive speech. It supports Chinese, English, Japanese, Spanish, and Arabic, with reference-audio cloning, emotion references, emotion vectors, and text-directed emotion.
VoiceStudio runs IndexTTS in a dedicated subprocess and Python environment.
This keeps its transformers<5 dependency isolated from VoiceStudio's
runtime. Existing user-managed IndexTTS-2 environments remain supported.
Install
IndexTTS 2.5 is not bundled because its source environment and model weights require substantial disk space.
- Open Model Catalogue → Engines.
- Expand IndexTTS 2.5 and select Install.
- Keep VoiceStudio open while source, dependencies, and weights download.
The installer:
- checks for
uvand at least 12 GB of free space; - installs the reviewed
indextts-2.5source revision in an isolated venv; - downloads the reviewed
IndexTeam/IndexTTS-2.5model revision; - resumes partial model downloads;
- saves
OMNIVOICE_INDEXTTS_DIRand activates the engine without a restart.
An app-managed IndexTTS-2 checkout remains intact while 2.5 installs into a
separate directory. VoiceStudio switches to 2.5 only after the new source,
environment, and weights pass verification. User-managed clones are never
modified or removed; their legacy
indextts.infer_v2 entry point remains supported.
Manual install
Use a separate checkout and venv. Do not install IndexTTS into VoiceStudio's root environment.
git clone --branch indextts-2.5 https://github.com/index-tts/index-tts.git
cd index-tts
uv venv .venv
uv pip install --python .venv/bin/python -e .
hf download IndexTeam/IndexTTS-2.5 --local-dir=checkpoints
On Windows, replace .venv/bin/python with .venv\Scripts\python.exe.
Then set OMNIVOICE_INDEXTTS_DIR to the checkout root:
export OMNIVOICE_INDEXTTS_DIR=/path/to/index-tts
[Environment]::SetEnvironmentVariable(
"OMNIVOICE_INDEXTTS_DIR",
"$env:USERPROFILE\code\index-tts",
"User"
)
Restart VoiceStudio after setting a persistent environment variable outside the app.
Compatibility
VoiceStudio probes these locations in order:
${OMNIVOICE_INDEXTTS_DIR}/.venv/;backend/engines/indextts/.venv/;- a venv bootstrapped from
OMNIVOICE_INDEXTTS_DIR.
The probe prefers indextts.infer_v2_5 and falls back to
indextts.infer_v2. A timed-out import is treated as unproven rather than
missing, preventing slow disks or antivirus scans from hiding a valid venv.
Set OMNIVOICE_INDEXTTS_IMPORT_PROBE_TIMEOUT_S to raise the default 60-second
probe limit.
Long-text generation
A long passage can keep infer() busy for several minutes. The sidecar emits a
keep-alive frame every 5 seconds while it works, so the parent can tell a slow
synthesis from a wedged one, and waits up to 900 seconds for a sidecar that has
gone genuinely silent. Set OMNIVOICE_INDEXTTS_RECV_TIMEOUT_S (minimum 30) to
tune that ceiling.
IndexTTS 2.5 requires a language token. VoiceStudio maps locale codes and language names to the five supported languages and detects Chinese, Japanese, or Arabic script for Auto requests. Ambiguous Latin text defaults to English.
IndexTTS 2.5 uses duration_factor for native duration guidance. VoiceStudio's
dubbing fit stage remains responsible for exact segment timing. Legacy
IndexTTS-2 installations continue receiving their target_tokens control.
Troubleshooting
Engine unavailable
Use Model Catalogue → Engines → IndexTTS 2.5 → Install. For a manual install, confirm that the configured directory contains:
pyproject.toml
indextts/infer_v2_5.py
checkpoints/config.yaml
IndexTeam/IndexTTS-2.5 ships the model config as config.yaml. Earlier
installs only worked after hand-renaming it to config_v2_5.yaml; both names
are accepted, so a renamed checkout keeps working as-is and needs no
reinstall.
uv not found
Install uv from https://docs.astral.sh/uv/ or configure the bundled binary
through OMNIVOICE_BUNDLED_UV.
Import fails after installation
For an app-managed install, retry Install to repair the source and venv. For a manual install, run:
uv pip install --python .venv/bin/python -e .
Insufficient disk space
Free the amount reported by the installer, then retry. Completed model files are reused.
License
IndexTTS 2.5 uses the bilibili Model Use License. It grants a limited, worldwide, non-exclusive, royalty-free license subject to its restrictions. A separate license is required when the user or an affiliate exceeded 100 million monthly active users in the preceding month or RMB 1 billion in annual revenue in the preceding year. The agreement also includes downstream, derivative-work, prohibited-use, attribution, and compliance obligations. Review the official license before installing or using the model.
This model is not covered by VoiceStudio's blanket commercial-use statement. Organizations above either threshold must obtain Bilibili's separate written license before installing or using IndexTTS 2.5. Other engines remain available without enabling this optional sidecar.
See Engine venvs and disk usage for storage details.