feat(voice): add Voicebox TTS provider for local voice cloning

[penso]

Summary

• Add Voicebox as an opt-in TTS provider behind a voicebox cargo feature flag
• Voicebox is a local Qwen3-TTS server with a FastAPI REST API for voice cloning
• Two-step generation: POST /generate returns metadata, then GET /audio/{id} fetches WAV bytes
• Feature flag chain: cli → gateway → voice, provider code gated with #[cfg(feature = "voicebox")]
• Config schema always parsed (no feature gate), matching existing provider patterns

Files changed

File	Change
crates/voice/src/tts/voicebox.rs	New TTS provider implementation
crates/voice/Cargo.toml	voicebox = [] feature
crates/voice/src/tts/mod.rs	cfg-gated module + re-export
crates/voice/src/config.rs	VoiceboxTtsConfig, TtsProviderId::Voicebox
crates/voice/src/lib.rs	cfg-gated re-exports
crates/gateway/Cargo.toml	voicebox feature forwarding
crates/gateway/src/voice.rs	Provider creation, listing, config loading
crates/cli/Cargo.toml	voicebox feature forwarding
crates/config/src/schema.rs	VoiceVoiceboxTtsConfig struct
crates/config/src/validate.rs	Schema map + semantic validation

Config example

[voice.tts]
provider = "voicebox"

[voice.tts.voicebox]
endpoint = "http://localhost:8000"
profile_id = "abc-123-def"
model_size = "1.7B"
language = "en"

Validation

Completed

• cargo +nightly-2025-11-30 fmt --all -- --check
• cargo clippy --workspace --all-targets (no warnings)
• cargo clippy --workspace --all-targets --features voicebox (no warnings)
• cargo check (default features)
• cargo check --features voicebox
• cargo test -p moltis-voice (124 passed)
• cargo test -p moltis-voice --features voicebox (130 passed)
• cargo test -p moltis-config (98 passed, schema drift guard OK)
• cargo test -p moltis-gateway --features voicebox (488 passed)
• taplo fmt

Remaining

• ./scripts/local-validate.sh

Manual QA

1. Enable feature: cargo build --features voicebox
2. Start voicebox server: voicebox serve
3. Set config: provider = "voicebox" in moltis.toml
4. Test TTS synthesis through the web UI or API

[codspeed-hq]

Merging this PR will improve performance by 96.29%

⚡ 1 improved benchmark
✅ 38 untouched benchmarks
⏩ 5 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	env_substitution	21.1 µs	10.7 µs	+96.29%

---

_{Comparing voicebox (0dd5a08) with main (5f2c245)²}

Footnotes

1. 5 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (9804455) during the generation of this report, so 5f2c245 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[codecov]

Codecov Report

❌ Patch coverage is 84.02778% with 23 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
crates/voice/src/tts/voicebox.rs	84.44%	14 Missing ⚠️
crates/gateway/src/voice.rs	72.22%	5 Missing ⚠️
crates/voice/src/config.rs	76.47%	4 Missing ⚠️

📢 Thoughts on this report? Let us know!