The meaningful architectural distinction with TTS-2 is that it operates as a closed-loop system. The model takes the actual audio of the prior turns of the exchange as input, not just a transcript \u2014 it hears how the user actually sounded. That\u2019s a non-trivial difference. A transcript of \u201cokay, fine\u201d gives you the words. The audio of \u201cokay, fine\u201d tells you whether the person is relieved, resigned, or sarcastic. TTS-2 is designed to use that signal.<\/p>\n
The same line lands differently after a joke than after bad news, and the model knows the difference because it heard the prior turn. Tone, pacing, and emotional state carry forward automatically. Practically speaking, audio context flows across turns inside a Realtime session without developers needing to pass explicit Inworld team is shipping TTS-2 with four key features,<\/strong> positioning the combination and not any individual piece, as the differentiation.<\/p>\n Beyond the four key features, it calls out a set of behaviors that push speech further into what it describes as \u201cperson paying attention\u201d territory. The most technically interesting is disfluencies: the model generates natural uh<\/em> and um<\/em>, self-corrections, mid-noun-phrase pauses, and trailing thoughts that signal warmth and recall rather than malfunction. Critically, different speaker profiles cluster fillers differently, and the model follows the rhythm \u2014 filler-as-energy sounds different from filler-as-hesitation.<\/strong> Voice cloning is also supported via a two-step API:<\/strong> upload a reference sample (5\u201315 seconds, clean, single speaker) to TTS-2 is one layer in Inworld\u2019s broader Realtime API pipeline. The full stack includes Realtime STT, which transcribes and profiles the speaker in one pass \u2014 capturing age, accent, pitch, vocal style, emotional tone, and pacing as structured signals on the same connection. A Realtime Router that routes across 200+ models, selecting<\/strong> the appropriate model and tools based on the user\u2019s state and conversation context. And TTS-2 at the output layer. The pipeline runs over a single persistent WebSocket connection, with sub-200ms median time-to-first-audio for the TTS layer.<\/p>\n Realtime TTS 1.5 already ranks #1 on the Artificial Analysis Speech Arena<\/a> (as of May 5, 2026), ahead of Google (#2) and ElevenLabs (#3). The launch of TTS-2 signals that Inworld considers raw audio quality a solved problem \u2014 and is now competing on the behavioral layer: context-awareness, steerability, and identity consistency across languages.<\/p>\n Check out\u00a0the\u00a0Docs<\/a>\u00a0<\/strong>and\u00a0Technical details<\/a><\/strong>.\u00a0<\/strong>Also,\u00a0feel free to follow us on\u00a0Twitter<\/mark><\/a><\/strong>\u00a0and don\u2019t forget to join our\u00a0130k+ ML SubReddit<\/a><\/strong>\u00a0and Subscribe to\u00a0our Newsletter<\/a><\/strong>. Wait! are you on telegram?\u00a0now you can join us on telegram as well.<\/a><\/strong><\/p>\n Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.?\u00a0Connect with us<\/mark><\/a><\/strong><\/p>\n The post Inworld AI Launches Realtime TTS-2: A Closed-Loop Voice Model That Adapts to How You Actually Talk<\/a> appeared first on MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":" Voice AI has a dirty secret: m…<\/p>\n","protected":false},"author":1,"featured_media":859,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-858","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/858","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=858"}],"version-history":[{"count":0,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/posts\/858\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=\/wp\/v2\/media\/859"}],"wp:attachment":[{"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=858"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=858"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/connectword.dpdns.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=858"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}prior_audio<\/code> fields or build additional plumbing.<\/p>\nFour Capabilities, One Model<\/strong><\/h3>\n
\n
[sad]<\/code> or [excited]<\/code>, developers pass a bracket tag like [speak sadly, as if something bad just happened]<\/code> directly in the text. Long, descriptive prompts beat short labels \u2014 the model responds far better to full context than single-word labels. Inline non-verbal markers like [laugh]<\/code>, [sigh]<\/code>, [breathe]<\/code>, [clear_throat]<\/code>, and [cough]<\/code> can be dropped anywhere in the text where the moment should occur, and the model places them as audio events, not pronounced words.<\/li>\nThe Conversational Layer Underneath<\/strong><\/h3>\n
\/voices\/v1\/voices:clone<\/code>, get a voice ID, and use it like any other voice.<\/p>\nWhere It Fits in the Stack<\/strong><\/h3>\n
![\"\"](\"https:\/\/www.marktechpost.com\/wp-content\/uploads\/2026\/05\/Screenshot-2026-05-05-at-5.20.25-PM-1024x377.png\")
The Broader Context<\/strong><\/h3>\n
\n