fix: word-by-word payload merging and browser client sample rate ...

- Вставка пробела между последовательными text payload'ами,
  чтобы слова не слипались при потоковой отправке (server.py:152-157)
- Браузерный клиент использует sample_rate из сообщения вместо
  хардкода 44100 (client_browser.html:125,208)
- AGENTS.md: добавлены команды браузерного клиента, gotcha
  про пробел между payload'ами
- docs/05_usage.md: раздел браузерного клиента переписан
  (HTTP-сервер, динамический sample_rate)

fix(server): use default voice_ref from settings when client omits it ...

Without this, browser/Python clients that don't send voice_ref failed
with 'F5-TTS requires a reference audio file' even when DEFAULT_VOICE_REF
was configured in .env.

chore(config): enable default voice ref, ref text and model name in .env.example ...

- .env.example: set DEFAULT_VOICE_REF, DEFAULT_REF_TEXT, TTS_MODEL_NAME
- config.py: add tts_model_name setting
- server.py: pass model name to F5TTSEngine constructor

Add explicit reference text config and clean default reference ...

- Add DEFAULT_REF_TEXT setting to skip Whisper transcription

- Pass ref_text through F5 backend and warm-up path

- Document new env variables and warm-up usage

- Update latency notes with explicit ref_text timing

Initial voice TTS server with F5-TTS backend, WebSocket streaming and warm-up ...

- FastAPI WebSocket API for real-time chunked TTS

- F5-TTS GPU backend with reference caching and auto-transcription

- Dummy backend for offline tests

- Streaming text segmenter, session state, stop/resume

- Startup warm-up, PCM16/base64 audio, basic tests

- Documentation: architecture, protocol, usage, roadmap, technical notes