Auphonic
Tool · this deep dive is built from the official pages of the platform itself and of the tools that support it.
Overview
Auphonic is automatic post-production for speech audio and video. The features page names: Intelligent Leveler (loudness between speakers and between speech and music), normalisation to loudness standards, noise and reverb reduction with separate control of noise, reverb and breath, filtering and AutoEQ with de-esser and de-plosive, bandwidth extension, cutting of filler words, coughs and silence, multitrack productions, speech recognition with shownotes and chapters, video, audiograms, subtitle burn-in, an editor, API, CLI and watch folders.
Input languages are known only in part. Of the recognition it is said verbatim: "Auphonic uses a multilingual Whisper model by OpenAI as self-hosted speech recognition engine"; Amazon Transcribe, Google Cloud Speech API, wit.ai and Speechmatics are connected as well. What is known is by plan: in the table on the pricing page the rows "Multilingual speech recognition" and "Automatic shownotes and chapters" stand at No on the free plan and at Yes on the paid ones.
Which platforms it works with
The result the tool publishes itself. Verbatim from the workflows page: "you can send your audio files to your favorite podcasting, audio, and video services such as YouTube, Libsyn, Blubrry, Facebook, Spreaker, PodBean, Podlove, Soundcloud, and many more". Metadata and chapters are exported to Soundcloud, YouTube and Spreaker. Also named are file transfer services, watch folders, Zapier, API and CLI, White Label API; a platform not on the list is connected through the API.
Pricing
Prices are from the pricing page, all excluding VAT, in two currencies. Recurring Credits paid monthly: S — 9 hours a month for 12 EUR or 13 USD; M — 21 hours for 28 or 30; L — 45 hours for 58 or 64; XL — 100 hours for 117 or 136; XXL — 250 hours for 250 EUR or 290 USD. Above 1000 hours a month it is Business by arrangement.
With annual payment "Pay yearly - Save 20%" is declared: S — 10 EUR or 11 USD a month; M — 23 or 25; L — 48 or 52; XL — 97 or 113; XXL — 210 EUR or 245 USD. The lowest subscription step is 10 EUR (11 USD) for S.
There are also one-off One-Time Credits packs — eleven steps from 5 hours for 11 EUR (12 USD) to 3000 hours for 2520 EUR (2925 USD).
What is free
Access is permanent and measured in processing time: "Auphonic is free for 2 hours of processed audio per month" — verbatim from the pricing page. Credits do not accumulate: "Free credits do not stack up if you don't use them". Into the output an audio insert is embedded: "Productions of free users come with a jingle".
What does not work here is on the same page: multitrack productions only under 20 minutes; multilingual recognition, automatic shownotes and chapters, watch folders and batch productions are unavailable; support is basic.
Restrictions
The length of the submitted material is counted in milliseconds, but with a floor per production: "we charge at least 3min of credits per production to avoid spam attacks".
Credits of three kinds behave differently. Free ones — 2 hours a month, not accumulating. Recurring ones reset monthly and do not carry over. One-Time ones are perpetual — "One-Time Credits never expire" — and are spent after the recurring ones but before the free ones.
By the terms the finished work is not kept long: productions are deleted after 21 days for audio and after 7 days for video and API.
Rights to the output
The rights stay with you, and the plan does not affect this. The terms say verbatim: "you fully retain any rights that you have with regards to your Content, including copyright". Restrictions on commercial use of the output are not in the document, and neither are differences by plan: the difference of the free step is not legal but audible — the row "Includes Auphonic jingle" stands at Yes for it alone.
Over everything a prohibition operates: the tool, its output and derivatives may not be applied "to develop, train, evaluate, benchmark, or improve any model, algorithm, or system", nor may competitors be built on them.
Do they train on your data
They do, and it is declared directly. The terms allow that "your Content may be viewed and/or listened to by an Auphonic employee or any person contracted by Auphonic to work on our audio processing algorithms". Individual fragments may be stored indefinitely to improve the algorithms — without access for the user and without a stated term.
The features page describes the training as ongoing: "Based on over five years of training with audio files from our web service, the algorithm keeps learning and adapting to new data every day". A way to opt out is not in what was read: neither a setting, nor a procedure, nor an address for a request is named by the terms.
How you earn with it
Commercial use of the output the terms do not restrict: the rights to the content stay with the user, there are no reservations about personal and non-commercial use, and by plan the rights are not divided. A processed podcast may be used exactly as the original recording.
The difference by plan is technical: into the free output a jingle is embedded, and it is removed by moving to a paid step, not by a setting.
One application is closed directly: training any models and algorithms on Auphonic's output and building a competing product on it is forbidden.
What it will not do
Voice it neither synthesises nor clones: such functions are not in what was read, the tool processes recordings already made. Consent from the person whose voice is recorded it will not require either. Special rules about someone else's voice are not in the terms: neither a requirement to obtain consent, nor a ban on processing the voice of a public figure, nor a description of a check. There are only general warranties — that the content infringes no third-party rights, contains no malicious code and breaches no agreement with an employer. A list of forbidden topics the document does not put forward either.
Keep the finished work long it will not — audio 21 days, video and API 7 days. Let you opt out of your audio being heard by employees and contractors it will not: an opt-out mechanism the terms do not describe.
Mark the result as machine-processed it does not promise: no watermark, no metadata, no C2PA is in what was read, and the jingle is a brand insert, not a note about the use of AI.
Verified data
The checked data this deep dive rests on.
About the platform
Automatic post-production of speech audio and video: Intelligent Leveler (balancing loudness between speakers, speech and music), loudness normalisation to standards, noise and reverb reduction with separate control of noise, reverb and breath, filtering and AutoEQ (de-esser, de-plosive), bandwidth extension, cutting filler words, coughs and silence, multitrack productions, speech recognition with automatic shownotes and chapters, video support, audiograms, burning in subtitles, an audio editor, an API, a CLI and watch folders
The list was checked against the site navigation and the feature comparison table on the pricing page
source, checked 2026-08-07
Speech recognition is multilingual: «Auphonic uses a multilingual Whisper model by OpenAI as self-hosted speech recognition engine» (verbatim); external services Amazon Transcribe, Google Cloud Speech API, wit.ai and Speechmatics are additionally connected. There is no explicit list of languages on the pages read, and no separate confirmation for Russian, Arabic, Hindi or Chinese either. DIFFERS BY PLAN: in the comparison table «Multilingual speech recognition» and «Automatic shownotes and chapters» are No on the free plan and Yes on the paid ones
The difference between plans is taken from the Feature comparison table at https://auphonic.com/pricing
source, checked 2026-08-07
Platforms
Automatic publishing of the result directly: «you can send your audio files to your favorite podcasting, audio, and video services such as YouTube, Libsyn, Blubrry, Facebook, Spreaker, PodBean, Podlove, Soundcloud, and many more» (verbatim). Metadata and chapters are exported to Soundcloud, YouTube and Spreaker. In addition: file transfer services, watch folders, Zapier, the Auphonic API and CLI, a White Label API; a third-party platform can be connected as a publishing target by the user through the API
The list of platforms is repeated in the Automatic Content Deployment section at https://auphonic.com/features
source, checked 2026-08-07
Pricing
Auphonic Free: «Auphonic is free for 2 hours of processed audio per month» (verbatim), credits do not accumulate («Free credits do not stack up if you don't use them»). A jingle is embedded in the result: «Productions of free users come with a jingle» (verbatim). Limits of the free plan: multitrack productions only shorter than 20 minutes; multilingual speech recognition, automatic shownotes and chapters, watch folders and batch productions are unavailable; support is basic
All the limits are taken from the pricing block and the comparison table on the same page
source, checked 2026-08-07
Recurring Credits paid monthly: S — 9 h/month for 12 EUR or 13 USD; M — 21 h for 28 EUR or 30 USD; L — 45 h for 58 EUR or 64 USD; XL — 100 h for 117 EUR or 136 USD; XXL — 250 h for 250 EUR or 290 USD; above 1000 h/month — Business by arrangement. Paid yearly (advertised as «Pay yearly - Save 20%»): S — 10 EUR or 11 USD a month; M — 23 EUR or 25 USD; L — 48 EUR or 52 USD; XL — 97 EUR or 113 USD; XXL — 210 EUR or 245 USD. One-Time Credits: 5 h — 11 EUR or 12 USD; 10 h — 21 or 23; 25 h — 49 or 55; 50 h — 86 or 100; 100 h — 147 or 171; 250 h — 275 or 320; 500 h — 470 or 545; 1000 h — 900 or 1045; 2000 h — 1740 or 2020; 3000 h — 2520 or 2925; above 3000 h — on request. All prices exclude VAT
The amounts were read from the data-price-euro and data-price-dollar attributes in the markup of the pricing tables: visually a script inserts them, but the values are in the HTML itself
source, checked 2026-08-07
10 EUR (11 USD) a month — the Auphonic S plan paid yearly; paid monthly the same S costs 12 EUR (13 USD) a month. The cheapest one-time purchase is a 5-hour pack for 11 EUR (12 USD). Prices exclude VAT
Both currencies are present in the markup; the EUR/USD switch merely selects which value to display
source, checked 2026-08-07
Limits and restrictions
What is counted is the duration of the material submitted for processing, in milliseconds, with a minimum of 3 minutes per production: «we charge at least 3min of credits per production to avoid spam attacks» (verbatim). Two hours a month are free and do not accumulate. Recurring Credits reset every month and do not carry over; One-Time Credits are permanent («One-Time Credits never expire») and are spent after the recurring ones but before the free ones. There is auto top-up when the balance falls below 1-10 hours. Re-running the same production with different settings is not charged; changing the input files is charged again. If processing fails, no credits are spent; if the result is unsatisfactory, they are refunded on request. Storage: productions are deleted after 21 days (audio) and 7 days (video and API)
The retention periods for productions are taken from the Terms of Service at https://auphonic.com/terms_of_service
source, checked 2026-08-07
Restrictions
WHOSE VOICE: THERE ARE NO SPECIAL RULES ABOUT SOMEONE ELSE'S VOICE IN WHAT WAS READ. Auphonic processes and restores recordings that already exist; no voice cloning or speech synthesis features are described on the pages read, so the terms contain neither a requirement to obtain consent from the person recorded, nor a ban on processing the voice of a public figure, nor a description of any check. There are only the user's general warranties: the content does not infringe third-party intellectual property rights, contains no malicious code and does not breach agreements with an employer. One separate ban: using the service, its output and derivatives to develop, train, evaluate, benchmark or improve any models, algorithms and systems, or for competing products. THERE IS NO list of forbidden content topics in what was read
The absence of voice rules is recorded from the Terms of Service and the features page at https://auphonic.com/features
source, checked 2026-08-07
Legal
Rights stay with the user: «you fully retain any rights that you have with regards to your Content, including copyright» (verbatim). The terms place no restriction on commercial use of the output, and there is no difference between plans in rights either. The practical difference of the free plan is technical rather than legal — an Auphonic jingle is embedded in the output. The user is forbidden to use the service, its output and derivatives «to develop, train, evaluate, benchmark, or improve any model, algorithm, or system» (verbatim) or to build competing products on them
The jingle being present only on the free plan is confirmed by the Includes Auphonic jingle row in the comparison table at https://auphonic.com/pricing
source, checked 2026-08-07
YES, training on user recordings is stated. Terms of Service: «your Content may be viewed and/or listened to by an Auphonic employee or any person contracted by Auphonic to work on our audio processing algorithms» (verbatim); individual audio fragments may be stored indefinitely for the sake of improving the algorithms, with no access for the user and no guaranteed term. The features page describes continuing training directly: «Based on over five years of training with audio files from our web service, the algorithm keeps learning and adapting to new data every day» (verbatim)
A second source is the Intelligent Leveler section at https://auphonic.com/features
source, checked 2026-08-07
There is no AI labelling: the Terms of Service and the pricing page contain no mention of a watermark, metadata or C2PA marking machine processing. A difference between plans does exist, but of another kind — an audio insert: in the comparison table the row «Includes Auphonic jingle» is Yes for Free and No for Monthly and One-Time and for Yearly and Business. This is a brand jingle, not a notice that AI was used
Marked partial because the difference between plans is confirmed, while AI labelling as such is not described in what was read
source, checked 2026-08-07
Catalogue section: all similar See also: catalogue index · scheduler comparison · find by situation · platform restrictions