Paste a paragraph or a 50-page doc. Long text is split into parts and starts playing while the rest generates. Markdown stripped automatically. No signup. Pick a realistic cloud voice, or one that runs entirely on your machine.
Also a Cursor & VS Code extension: Agent Auto-read speaks your agent's replies. One install, every window, zero echo. MIT on GitHub.
Cloud voices are generated by Microsoft, then dropped. Prefer that nothing leaves your machine? Use offline voices below. Privacy
Varterm is not a single voice. Different text wants a different trade-off — keeping it secret, getting it read now, or sounding convincingly human for an hour — so the same controls drive more than one engine.
Piper voices synthesise inside the tab. One model download of about 63MB is cached on the device, and after that there is no network call at all — read an unreleased spec or a contract and nothing leaves your machine. No round trip also means it starts fast, and keeps working on a plane.
Microsoft neural voices, unmetered and with no account. Genuinely realistic across American, British, and Australian accents, with nothing to download first. Your text goes out only to be turned into audio and is dropped immediately — never stored, never sold, never trained on.
An ElevenLabs integration for the most lifelike delivery we can offer, for the times a voice has to hold your attention through a long document rather than just get the words out. It bills per character upstream, so it will be a paid add-on. The two engines above stay free.
Drop in long articles, RFCs, and full chapters. Free cloud neural voices, chunked playback, live part progress.
Agent Auto-read speaks finished replies while you keep coding. Long answers split into parts you can jump through. One install, every window, zero echo.
Read a selection or an entire long-form page without leaving the tab you are on.
Send the prompt and go back to your file. When the agent finishes, the reply reads itself — you stop parking on the chat panel watching tokens stream in.
Long replies split into parts. Jump forward past the preamble, jump back to re-hear the one line that mattered. No scrolling up through the chat to find it again.
Run a fleet of Cursor windows across repos. One user install covers all of them and only the focused window speaks, so parallel agents never talk over each other.
Flip it on in Cursor and finished agent replies just play. One install covers every window and only the focused one speaks, so ten windows never echo. Play, pause, stop, and jump live in the status bar. Cursor text to speech extension.
Long articles, RFCs, and 40-page agent dumps. Varterm splits text into parts, plays part one while the rest still generates, and keeps going — no single-request timeout wall. Watch it track part 3 of 12 as it reads. Long-form text to speech guide.
Paste content from ChatGPT, GitHub, or any markdown source. Our TTS strips headers, links, code blocks, and formatting for clean, natural-sounding audio output. Markdown to speech guide.
No account, no paywall, no trial. The editor extension is MIT licensed and the source is on GitHub — read it, fork it, build the VSIX yourself.
There is no account, so there is nothing to profile. What you paste goes to the voice engine, comes back as audio, and is gone — no database, no history, no training set. We count page visits, never your words.
Microsoft neural voices that sound natural and expressive, in American, British, and Australian accents with adjustable speed — free and unmetered. Studio-grade ElevenLabs voices are coming as a paid add-on for when you need more.
Switch the web reader to Piper voices and synthesis happens on your machine. One 63MB model download, cached on the device, and after that your text never leaves the tab — no server, no network, nothing to leak.
No waiting in queues. Audio generates instantly and starts playing immediately with reliable long-form playback.
Yes. Varterm is free to use with no hidden charges and no registration. You get unlimited access to Microsoft neural cloud voices, offline Piper voices that run on your own machine, and your browser voices. A studio-grade ElevenLabs tier is planned as an optional paid add-on, and the existing voices stay free.
Varterm supports long form text to speech conversion for documents of any length. Large texts are automatically split into chunks and played sequentially for seamless listening. There are no character limits.
Yes. Enable the Strip Markdown option to remove headers, bold, links, code blocks, and other formatting for cleaner speech output. It works well for ChatGPT answers, GitHub READMEs, and docs.
No account or registration is needed. Paste your text and click play. Text is processed to generate audio, and we use basic analytics to improve site reliability.
Varterm runs more than one engine behind the same player. Cloud voices use Microsoft neural voices for realistic speech with nothing to download. Offline Piper voices run locally in your browser, so your text never leaves the device. Your device built-in browser voices are also available. Studio-grade ElevenLabs voices are planned as a paid add-on.
Choose offline Piper voices when the text is confidential or you are working without a network, since synthesis happens entirely on your device. Choose cloud neural voices for the most natural free quality with no model download. Both are free; a paid studio tier using ElevenLabs is planned for the most lifelike delivery.
Yes. Paste articles, documentation, or ebooks of any length. Varterm chunks long text automatically for smooth playback without character limits.