Speecho What changed in Speecho, newest first. The same updates are behind the "What's new" button in the app.
Twenty-four new voices on Enhanced and Premium speak three more languages of the European Union as natives: eight each in Bulgarian, Croatian and Slovak, four women and four men in every one. Pick the language in the voice list to hear them. Each language has its own page with samples, in English and in the language itself.
Hear all the voices →Open an EPUB, Word or PDF file in Text to speech: when it marks its chapters (a contents list, headings or bookmarks), every chapter becomes its own MP3, named after it. Untick the ones you do not need, like the contents page, and download the chapters as one ZIP. The price is the same as for any text.
Turn an ebook into an audiobook →List mode in Text to speech now makes flashcards. Open a CSV with words in one column and translations in the other, paste two columns from a spreadsheet, or write each card as word → translation: the words are voiced in one voice, the translations in a voice of their own language, and every card becomes one MP3 (the word, a pause of 0.5, 1 or 2 seconds, the translation) or two files, one per side. The ZIP comes with a CSV for Anki whose sound fields name the files. Up to 1,000 cards, priced as a list of all the sides.
Flashcard audio →Forty-three new voices on Enhanced and Premium speak six more languages as natives: eight in Czech and seven each in Greek, Hungarian, Norwegian, Danish and Finnish, with at least three women and three men in every one. Pick the language in the voice list to hear them. Each language has its own page with samples and a side-by-side with a universal voice.
Listen to the voices →Text to speech has a third mode, List: paste words or short phrases, one per line, or open a .txt file or the first column of a CSV, and every line is voiced as its own MP3 in one voice and one set of settings. The files are named after their lines and numbered in order, and the whole set downloads as one ZIP. Up to 1,000 lines per list. The list is priced as one text, with one minimum for all of it and every line counted as at least 10 characters, so a thousand short words cost about 1,600 credits on Premium. Any line can be played on its own, and chosen lines can be voiced again, or corrected first, each at its own characters.
Bulk text to speech →Until your first pack, free credits voice up to 300 characters per clip and 600 in total in Enhanced and Premium voices: enough to hear your own text in the voice you picked. Standard voices use the free credits with no such limit, and any pack removes it. Voice previews stay free.
Text to speech →Clips in Enhanced and Premium voices now download with subtitles, as SRT or VTT, from the arrow next to MP3. The timing follows the audio word by word, at any playback speed, so the file goes straight into a video editor; a dialogue can have the speakers' names in front of their lines. In a long text or a dialogue, one part or one line can now be voiced again on its own: in another voice, with a word fixed, or with a direction like [whispers], while the rest stays as it is.
Text to speech →In Text to speech, "Make a podcast from text" turns an article, a chapter or your notes into a conversation between two hosts, about 5, 10 or 20 minutes long, in the language of the text or another one you pick. The hosts stick to what the text says. The script opens in the Dialogue editor, where you check it, change the voices and voice it into one MP3. Writing the script costs 15 credits for up to 5 minutes, 25 for up to 10, 40 for up to 20, plus 1 credit per 3,000 characters of text; voicing it costs the same as any dialogue.
Open Text to speech →Live subtitles can now appear as the platform's own captions, with no bot in the call. Paste the caption link from Zoom (the API token, on a free account too) or YouTube Live (the captions ingestion URL) under Captions in a meeting, or connect OBS to put closed captions inside your stream. Send the original or the translation. Google Meet has no way to take outside captions. Needs an account; it costs the same as live subtitles.
Captions in Zoom and YouTube Live →Identify speakers now runs on a newer speech engine: fewer words on the wrong person, no skipped passages. The same engine now transcribes Armenian, Icelandic and Nepali, with several times fewer mistakes in our tests. Same price, nothing to switch on.
Transcribe a file →The live interpreter now has two tiers, chosen before each session. Standard reads the translation in one voice you pick, follows your glossary of names and terms, and costs 40 credits a minute: an hour of interpreting for the $5 pack. It runs about five seconds behind the speaker and covers 50+ languages. Premium is as before: several voices, about three seconds behind the speaker, 70+ languages, 80 credits a minute.
Live interpreter →In Conversation, turn on "Speak the translation aloud" and each turn is spoken to the other person in their language as soon as it ends, so nobody has to read the screen. Pick a voice for each of the two languages: the choice is remembered. With headphones the microphone stays open while a translation plays; without them it pauses, so the phone does not hear itself. With voice a conversation costs 40 credits a minute.
Open Conversation →Twenty-five new voices on Enhanced and Premium fill out the regional side of four languages: Mexican and Castilian Spanish now have seven voices each, Canadian French from Québec six, European Portuguese six and Flemish six. In the picker they carry a tag next to the name: MX, ES, Québec, PT, BE.
Listen to the voices →In Text to speech, switch to Dialogue and write lines for different characters, each with a voice of its own, then get one MP3 with a short pause between lines. A script pasted as "Anna: ..." splits into lines by itself, and a transcript with speakers opens as a dialogue. It costs the same as the text in Text mode.
Multi-voice text to speech →Captions on a stream now change as you adjust them: pick a color (any color), left, center or right alignment, text size, outline or a dark band, and the overlay in OBS follows without a new link. Captions fade out after a few seconds of silence, with the delay up to you. And there is a permanent overlay link for your account: paste it into OBS once and your sessions show there from then on, without making them public.
Live captions for OBS →Put the phone between two people who speak different languages and talk: each reads the other in their own language, on a split screen whose top half faces the other person. It works in 60 languages and tells who is speaking by the language, with hold to talk for noisy places and languages that sound alike. The conversation is saved in both languages and costs the same as live subtitles.
Conversation translator →Tell live subtitles the names and words of your session and how to translate them: a pastor's name spelled right, Kirchenschiff translated as nave rather than church ship. Names come from the same list you use to transcribe files, translations are kept per language, and a change applies mid-session too.
Open live subtitles →Recordings that came in pieces can now be read as one. In your history, press Select, tick two or more transcripts from the same tab and press Merge: they become one transcript with timestamps running on and speakers numbered across the parts, and the audio is joined too when every part kept it. Merging is free, the originals stay, and the same Select deletes several items at once.
Open your history →Standard now has native speakers of 18 languages next to its 13 universal voices, at the same 50 credits per 1,000 characters. They also read pause and emotion tags such as [pause] and [whispering].
Listen to the voices →Type a tag in square brackets to change how the next words sound: [whispering], [excited], [laughs], [slowly]. Enhanced and Premium also take accents, sound effects and any short direction of your own, and native Standard voices read a set list. The tags sit next to the text box, and a click adds one to your text.
See the tags →The voice picker now filters by gender, age and what a voice suits best: narration, conversation, ads and social, learning or characters. Your filters stay set for next time.
Open the voice picker →Speak, and people hear you translated into one of more than 70 languages a few seconds later, in a natural voice, with the text of the original and the translation on screen. Share a link or a QR code and up to 20 listeners follow on their own phones, with no app or account, or translate a Zoom, Meet or Teams call playing in a browser tab. Both texts are saved to your history, and Keep audio can keep a recording of the original.
See how it works →Enhanced and Premium voices now run on a newer voice model.
See the voices and prices →Files transcribed without speaker labels now go through a newer speech recognition engine. On real recordings it made clearly fewer mistakes in languages other than English, with names, and with speech that mixes languages. English is as accurate as before. Nothing to switch on, and the price stays the same.
Transcribe a file →Next to the speed buttons in the voice player there is now a field for any speed from 0.5× to 2×. The speed you pick applies to all your clips. The MP3 button saves a clip at normal speed, and the arrow next to it saves it at yours, a long text joined into one file included, with the voice at its normal pitch and no extra credits.
Open the Voice tab →English gained seven American, four British, three Indian, two Australian and two Irish voices, from short-form creator voices and a Southern storyteller to audiobook narrators. Spanish, German, French, Portuguese, Italian, Japanese, Arabic, Korean, Dutch, Polish, Turkish and Ukrainian each gained native voices too, picked from the most-used voices in the ElevenLabs library. Arabic now has a Moroccan voice, and the two Mexican Spanish voices carry their own tag. Every voice has a free sample.
Hear all 203 voices →Sign in with a passkey using your fingerprint, face or device PIN, or turn on an authenticator app with one-time recovery codes. You can also change your password and email address from the app, see every device signed in and sign any of them out on its own, choose how long transcripts are kept, and download everything your account holds as a zip.
Open account settings →Every top-up and refund is now listed in your account with its transaction id, and each purchase offers the official Paddle invoice as a PDF whenever you need it. A new Usage tab charts what your balance was spent on day by day, and business buyers can add a company name or VAT id to their invoices.
See your payments →A pause goes in as [pause] or [long pause] on the Enhanced and Premium voices, and a word in CAPITALS gets extra stress. The Voice tab lists all three next to the text box, and the text-to-speech page explains them.
See how it works →Twenty-one new voices: Australian, Indian, Irish and Scottish English, plus English read by German, French, Italian and Japanese speakers - for characters, dubbing and language teaching. Each accent has its own page with free samples, and the English voice page now links them all.
Hear the accents →The German, Ukrainian and Arabic voice pages now exist written in those languages, for people who look for a voice in their own language rather than in English - the Arabic one reads right to left. Live translation also gained English to Vietnamese, German to Arabic and Korean to English.
See the German page →The German-to-Ukrainian and English-to-Spanish pages now also exist written in German and in Spanish, for the person organising the event rather than for an English-speaking reader. Both open the app with the pair already selected.
See the German page →The six US-accent voices and the four UK-accent ones now each have a page with their own samples and a side-by-side of the two accents, so you can hear which one your audience will read as local. The text-to-speech hub also lists every language with how many native voices and which accents it has.
Hear the US voices →Tamil to English, Polish to Ukrainian, German to Ukrainian and English to Arabic now have their own pages, each opening a live session with both languages already selected. The Arabic subtitles are set right to left, in the viewer link and in the exported transcript. Nine pairs have a page now; every other combination still works the same way from the app.
See the language pairs →Drop a document into the Scan tab and its text lands in your history for free: it is read in your browser, so no credits are spent. From there it can be translated, summarized or voiced. Scanned PDFs without a text layer can still go through paid OCR, up to 30 pages at a time.
Open the Scan tab →Paste a whole article or send a long transcript to the voice tab, up to 200,000 characters. It is voiced one part at a time: hear a short sample or part 1 before you pay for the rest, and stop whenever you like.
Open the voice tab →Thirty-two native speakers joined the voice catalogue. Open the voice picker and see whether your language is one of them.
Hear the new voices →Five pages set up a live session for a specific pair - English to Spanish, Spanish to English, English to Polish, German to Vietnamese and Portuguese to English - and open the app with both languages already selected. Any other combination still works the same way from the app.
See the language pairs →A saved live session can be picked up again from your history or straight from the ended session. The new part is added to the same transcript, the same audio file and the same share link, so a lecture that ran over or a meeting with a break stays one entry.
Live subtitles →Three new caption styles, Karaoke, Pop and Fill, light up each word as it is spoken, the way captions look on Reels and Shorts. You choose the highlight colour and how many words stay on screen at once; the preview animates a sample so you see the effect before you render.
Add captions to a video →Burned-in captions are now one step: drop a video, pick a style, render. Position, size and colours are previewed on a frame of your own video, with a Reels and TikTok safe-zone guide. The transcript is billed like any upload, the render by the second with a one-minute minimum, and finished videos stay downloadable for 30 days.
Add captions to a video →Live subtitles are tuned for speed. With Keep audio on, a finished session can be run through the full-accuracy transcription afterwards and you get a cleaner text next to the live one, billed like a file upload.
Live subtitles →An opt-in for uploads and live sessions: the recording is stored next to the transcript for 90 days, private to you. Play it back from any timestamp while reading, download it as mp3, or delete just the audio. Off by default, so nothing changes unless you ask for it.
Open the app →