AwesomeTTS vs HyperTTS
The comparison people expect is between two competing add-ons. It is not. The same person maintains both, one is the rewrite of the other, and the honest question is which interface you want and how the audio reaches your phone.
The two side by side
Checked against both AnkiWeb listings, both repositories and the author's own pages in August 2026.
| Aspect | AwesomeTTS | HyperTTS |
|---|---|---|
| Development in 2026 | Critical fixes only, by the author's own statement | Active, releases every few weeks |
| Relationship | The original, from 2012 | Rewritten from scratch by the same maintainer |
| Audio saved into the collection | Yes, batch generation | Yes, called Collection Audio |
| Audio spoken at review time | Yes, called on the fly | Yes, called Realtime |
| Cloze notes | Review-time audio only | Review-time audio only |
| Voice choice | Per service | All services in one filterable list, plus random and priority modes |
| Free voices without an account | Yes, and several have been breaking | Yes, with a switch that turns them all on |
| Paid tier | Same subscription, one key covers both | Same subscription, one key covers both |
Why there are two at all
AwesomeTTS was written by Arthur Helfstein Fragoso in 2012, went unmaintained around 2019, and was picked up by Luc Wastiaux, who has maintained it since roughly 2020. In 2022 he rewrote it from scratch rather than keep patching it, and called the result HyperTTS. The AwesomeTTS listing on AnkiWeb now opens by asking you to consider switching, and says plainly that AwesomeTTS will only receive critical bug fixes from here.
Searching AnkiWeb for AwesomeTTS returns three listings and two of them are dead: 301952613 was last updated in 2017 and works only with Anki 2.0, and 814349176 is an unofficial fork whose own description says it no longer works. The maintained one is 1436550454. HyperTTS is 111623432.
The paid plan is the same product under two names, HyperTTS Pro and AwesomeTTS Plus, and one key works in both add-ons. So choosing between them is a question of which interface you want, not of which company you pay.
The choice that actually matters
Both add-ons offer the same two ways of producing sound, under different names. Picking the wrong one is the mistake that costs you the voice you chose.
The add-on generates an mp3 for each note and writes a [sound:...] tag into a field, exactly as a shared deck would. The files live in your collection.media folder, so they sync to AnkiDroid and AnkiMobile, play offline, and keep the voice you paid for. This is the mode you want for vocabulary, and the one limitation worth knowing is that it does not work on cloze notes.
Instead of files, the add-on writes an Anki {{tts}} tag into the card template and speaks the field as the card is shown. Nothing is stored, so the collection stays small and edited text is spoken correctly straight away. It is the only route that voices cloze notes.
Because it is an ordinary template tag, it keeps working on your phone, but the phone has no add-on and no cloud key, so it falls back to whatever voice the operating system provides. The premium voice you picked on your computer is a desktop-only experience in this mode. HyperTTS's own tutorial states this: the tag is compatible with AnkiMobile and AnkiDroid, but the voice will be different.
If you study on a phone at all, generate files into the collection rather than relying on review-time speech, whichever add-on you use. The exception is cloze, where files are not an option and the OS voice is what you get on mobile.
What it costs
Google Translate, Naver Papago, DWDS for German and Youdao for English cover full sentences without an account. Oxford, Cambridge, SpanishDict and Duden cover single words. Windows voices, macOS voices and eSpeak NG run locally with nothing online at all. HyperTTS has a single switch that enables every free service at once.
Azure, Google Cloud, Amazon Polly, ElevenLabs, OpenAI, Forvo and the rest want an API key. You can bring your own from the provider, which is what heavy users do because the providers have free tiers of their own, or you can pay the add-on author's subscription and skip the provider accounts entirely.
Five dollars a month for 250,000 characters, which the author estimates at 3,000 to 6,000 notes. There is a seven-day trial capped at 50,000 characters. One key covers AwesomeTTS, HyperTTS and Language Tools.
The quota is not counted in plain characters for every service. Azure counts each Chinese, Japanese and Korean character as two, OpenAI counts two, Naver counts six, and ElevenLabs counts fifteen. That last one turns 250,000 characters into about 16,000 of actual text, which is the difference between a year of vocabulary and a fortnight of it.
What people complain about
From the reviews on both AnkiWeb listings, 2025 and 2026.
The recurring complaint through 2025 and 2026 is not the add-on itself but the services it scrapes. Naver Papago returned 404 errors across a run of reviews in mid-2026, and the Youdao dictionary was broken for months in early 2025 before being fixed. This is the structural risk of free voices that are somebody else's website.
The standout report in 2026 is presets disappearing when the add-on updates, which shows up repeatedly in reviews and which the author acknowledges as an open issue rather than dismissing. Worth knowing before you configure twenty note types.
Users migrating from AwesomeTTS describe the fixed source and target fields as inflexible: renaming a field can break a preset, and audio goes to a predefined field rather than to the cursor. If you liked AwesomeTTS's looser editor workflow, this is the change you will feel.
Several one-star reviews are about being asked to subscribe, and several five-star reviews are from people who read those and then set the add-on up for free without trouble. Both positions are accurate: the premium voices cost money, the free ones do not, and the add-on does ask.
Anki can already do some of this
Anki has spoken fields on its own since 2.1.20 with a {{tts}} tag in the card template, and it works on AnkiMobile and AnkiDroid too. It costs nothing, stores nothing, needs no key, and handles cloze. The voices are whatever the device has, which on Windows and macOS is decent and on Linux is nothing at all unless an add-on supplies it.
Both add-ons' review-time mode is a graphical front end to that same tag. What you are paying for is the other mode: real mp3 files, generated in bulk, in a cloud voice better than your phone's, stored in your collection.
Frequently asked questions
HyperTTS, unless you already have AwesomeTTS set up and working. It is the one being developed, and the author says so himself on the AwesomeTTS page. Migrating an existing, working setup for its own sake is not worth it.
The add-ons themselves are desktop only. Audio they save into your collection syncs and plays on AnkiDroid and AnkiMobile like any other media. Review-time speech keeps working on mobile but with the device's own voice.
Yes. Several free services cover whole sentences and the local system voices need no account at all. The subscription buys the premium neural voices without you having to open provider accounts.
No. A deck generated here arrives with the mp3 files already inside the .apkg, so it speaks in Anki on every device with nothing installed. Add-ons solve the other problem: a deck you already have that has no audio in it.
Yes, and that is exactly what they are for. Select the notes in the Browser, choose the source and target fields, pick a voice and generate. That is the case where an add-on beats any generator, because the cards already exist.
Keep reading
Related guides that build on what you just read.