feat(smallestai): add word_timestamps support to TTS

[harshitajain165]

Summary

Adds word_timestamps opt-in flag to the Smallest AI TTS plugin, matching the feature shipped in Lightning v3.1 and v3.1 Pro on 2026-06-01.

Changes

• Added word_timestamps: bool to _TTSOptions dataclass
• Added word_timestamps: bool = False parameter to TTS.__init__
• Added word_timestamps to update_options
• In SynthesizeStream._run_ws: includes word_timestamps: true in the WebSocket payload when enabled, and handles the new word_timestamp status frame

Usage

tts = smallestai.TTS(
    model="lightning_v3.1_pro",
    voice_id="meher",
    word_timestamps=True,
)

When enabled, the server interleaves status: "word_timestamp" frames with audio chunks. Each frame carries data: { id, word, start, end } where start/end are floats in seconds relative to the start of the audio stream.

Supported voices: base-queue English + Hindi voices (meher, devansh, kartik, maithili, liam, avery). Other voice families silently emit no word events — audio works normally.

Defaults to False — purely opt-in, no behavior change for existing integrations.

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 1 additional finding.