Omphalis adds voice conversations to source-grounded search

Voxiven LLC has launched voice conversations in the Ask feature of its Omphalis reading app, letting users talk through saved articles, papers, podcasts and videos while seeing numbered source passages behind each answer. The rollout is live now in the web app, with mobile support planned and voice limited to paid plans after a 30-day trial.

Why it matters: - Omphalis is trying to make reading and research feel more like a live conversation, while keeping every answer tied to the user’s chosen sources. - The feature is designed for people who want to understand dense material without leaving the page, opening a separate search tool or losing the thread of a question. - Visible citations and source passages make it easier to check whether an answer is supported, incomplete or wrong.

What happened: - Voxiven LLC launched voice conversations in Ask, the question-answering feature inside the Omphalis web app. - Readers can speak questions about saved articles, research papers, podcast episodes and videos, then hear answers drawn from that material. - Follow-up questions can also be asked aloud, keeping the conversation in the same thread. - The launch is live now in English in the Omphalis web app.

The details: - Ask can use the item currently open, a Space of related material or the reader’s full Library. - Ask proposes a starting scope from the question, and the reader can change that scope at any time. - A fourth option, Omphalis guide, answers questions about using the app itself. - Answers about a reader’s material draw only on that reader’s Library and never on anyone else’s. - Each spoken answer appears with numbered sources beside it. - Selecting a source opens the cited passage in the reader, highlighted within the surrounding text. - Ask does not read citations aloud. - If the selected sources do not cover a question, Ask is designed to say so rather than guess. - The voice experience is built for low latency, so Ask starts speaking as an answer takes shape instead of waiting for the full response. - Voice runs in the browser beside the text being discussed. - Pressing Talk starts a session. - Ask greets the reader by name, mentions what is open and then listens. - A live transcript shows each question as it is spoken. - Speech is transcribed and voiced by Deepgram. - Omphalis does not keep the audio. - A voice session lasts up to 20 minutes. - Each question and answer remains in the conversation panel as text, with sources, after the session ends. - Omphalis also brings web articles, research PDFs, EPUBs, Word documents, podcast episodes and YouTube videos into one reading environment. - Audio and video are turned into transcripts that can be read and questioned alongside written material. - The structure panel maps each piece into sections and recurring names and terms. - Dotted underlines in the text open short explanations of difficult terms and references. - Notebooks let readers keep passages beside their own writing. - A notebook’s Ask cells answer questions from its contents and keep cited responses and follow-up threads. - The company says Ask voice is meant for students, researchers, professionals and listeners returning to long material.

Between the lines: - Omphalis is positioning itself as a source-first alternative to generic AI chat tools by making the evidence visible instead of hidden inside the answer. - The product design suggests a bet that spoken interaction will help readers stay focused on the original text longer than typing into a separate interface. - The app’s combination of transcripts, notes, structure maps and cited answers points to a broader effort to turn passive reading into active study.

What's next: - Mobile support is planned. - New accounts get a 30-day trial of every feature, including voice, without a credit card. - After the trial, accounts move to Free unless a paid plan is chosen. - Voice is included in paid plans: Pro at US$9 per month, Premium at US$19 and Scholar at US$39. - Each paid tier includes a monthly voice allowance that increases with the plan. - The Free plan does not include voice. - More details are available on the Omphalis pricing page.

The bottom line: - Omphalis is betting that readers want an AI assistant that talks back, but only after showing its work.