STT plugin with voice profiling; TTS connection hardened

Patch Changes

• feat(inworld): add STT plugin with voice profiling support - #1516 (@karan-dhir)

  Adds STT and SpeechStream to the Inworld plugin, porting the Python livekit-plugins-inworld STT implementation to TypeScript. Supports both streaming (bidirectional WebSocket) and batch (REST) modes, word-level timestamps, and acoustic voice profiling (emotion, accent, age, pitch, vocal style) via SpeechData.metadata.voiceProfile.

• Harden the Inworld TTS connection layer (retryable framework errors, a handshake timeout and a bounded in-turn wait, and no in-plugin backoff so retries are delegated to the framework) and change the defaults to the inworld-tts-2 model and Jason voice. - #1675 (@cshape)

• Updated dependencies [278af00740f8ade253485afc1d51a19382cb5a2c, 7c1cfcb1237f4aa7d36669754fc388bbaa872a71, 703cffbfda8a8b830b78f2c89e0b2284244dc120, 22b24ae18ca4f3dd690a948e9f9ab3ac9cdad4b0, 8524731b39097fe521952f83b210b999834bbeda, aa8d427391a79ec8e2ccda103e1460c720b73acf, 4207d54c037ad0206814ae5e246735e357bb4c3b, 106c2b6820751ce3802d77e1e1e74ece130efc89, 710d112f70ce1134f3098729cda6a950d1deb66c, 7479bd4640d2d82a4f0263dc0745e72bab733127, b157d241ffaf402c0a2b8dc05dbf7e9ed96d21ee, 181c868d75fe8828e5193fc530f454d52fc776e0, f470e5531da74d66484194125fc0f950b5a53431, 453b69965e0075f634f79c323194ceb16ccfe1a6, 0a27a700dcef89e6202872024d73452b47025788, 0f29f6bd34559ba98544c2c9621936b44a0573b8, c5eaadd27748aa2a69a4cb1a72b3de652fa841ef, 154189e4af2149c580aeaba7dfe3a8c79f3d682d, d85a30757405158fb3182be1c77aaec6acecdd67, ee078ec394e697a989bd0faf53aa043392fa7ae4, 7af2f8ebafe86fae52854ddc73b253f2d29da286, 33578844ba169f1485465ea96346567c449b7ea7, 62c8278eb1294088cad66285a799e84e20799eaf, 35a33ff0f89a23b17ea3ad79c642188e66e31604, ee078ec394e697a989bd0faf53aa043392fa7ae4, dfb34c27b8cb075ed77e63ceb5286fc504122abd, 98a46dd366fa69f9270d3ae780e50740ffe72c8a, 278af00740f8ade253485afc1d51a19382cb5a2c]:

  • @livekit/agents@1.4.5