Proso por Pedro Balbino

<p>Proso reads web pages and selected text aloud with natural voices while keeping the text and its playback controls together on the page.</p>
<strong>What it does</strong>
<ul><li>Extracts the article text or reads a selection you highlight.</li><li>Starts from the popup, a paragraph you click, a keyboard shortcut, or the context menu.</li><li>Highlights the current sentence and word while the audio plays.</li><li>Switch voice mid-read from the player — no restarting the article.</li><li>Pause, resume, speed, seek, previous/next paragraph, and a listening queue for what piles up.</li><li>Caches generated audio locally and can export audio and highlights.</li><li>Works without a Proso account through BYOK or a synthesis host you operate.</li></ul>
<strong>Choose your audio route</strong>
<ul><li>Bring an OpenAI, ElevenLabs, Groq, or Cartesia API key.</li><li>Enter the address of a synthesis host you operate; Proso requests the narrowest Firefox host permission and sends text only to that exact configured destination.</li><li>Eligible plans may use Proso-managed synthesis.</li></ul>
<strong>Part of the Lectrice family</strong>
<p>Proso is the web member of the Lectrice family of reading tools, sharing its voices and its design.</p>
<p>Reading necessarily sends the text you ask to hear to the selected synthesis destination. Nothing is sent until you start reading. The public AMO build sends no usage telemetry, interaction analytics, browsing history, or page URLs to Proso.</p>
<p>Firefox 109 or later is required. An API key, configured synthesis host, or eligible managed plan is required to generate audio.</p>