A first-use test found a powerful automated finishing tool for loudness, noise, filler words and publishing. But creators still need to review its cuts, protect sensitive material and make the editorial decisions themselves.
Creator Tools | First Look
By Brandon L. Burley
Disclosure: Auphonic provided The Redemption Project with 20 hours of one-time promotional processing credits to test the service. Auphonic also offered TRP access to its 20% affiliate program. This article contains no affiliate links. The company did not request, review, edit or approve this article and had no say in its conclusions.
KNOXVILLE, Tenn. — The first Auphonic screen did not look like a full audio studio. It looked like a large upload box.
Drop in an audio or video file. Choose a single-track or multitrack production. Tell the system what technical work to perform, then let it analyze and process the recording.
That simplicity is much of the product’s appeal.
For my first test, I processed a roughly 47-minute episode of The Redemption Project, “Risk Assessment Under Fire,” featuring bail-reform advocate Ken W. Good. The finished-production screen placed the original and processed waveforms on top of one another while the editor displayed a time-aligned transcript, algorithm segments and controls for inspecting or changing cuts.
The processed waveform looked noticeably more consistent than the source.
That established that Auphonic had done technical work. It did not answer the more important editorial question: Did it make the right decisions?
After one production, this remains a first look rather than a final recommendation. But Auphonic does enough in one workflow to deserve attention from independent podcasters who spend too much time fixing uneven voices, background noise, pauses, filler words, loudness and publishing details after the real edit is finished.
Auphonic is not actually new
Auphonic may be new to my production workflow, but it is not a new startup.
The company opened its automated post-production web service to public users in 2012. Its current platform has expanded far beyond basic loudness normalization into automatic cutting, transcription, chapters, show notes, multitrack mixing, video processing and direct publishing.
In 2026, the company added automatic cutting for video, a beta “Studio Voice” system that reconstructs damaged or degraded speech and the ability to burn styled captions directly into video.
That broader capability makes Auphonic potentially more useful.
It also makes understanding its limits more important.
What Auphonic can do well on paper
Auphonic is best understood as an automated post-production system.
Its single-track algorithms analyze a finished mono or stereo recording and attempt to correct level differences among speakers, music and speech. Loudness normalization and a true-peak limiter can then prepare the final file for podcast, broadcast and other common loudness standards.
Multitrack processing gives the system more information.
When speakers are recorded separately, Auphonic can process individual tracks while building a final mix. Its documented tools include leveling, dynamic-range processing, gating, noise and hum reduction, microphone-bleed removal, ducking and filtering.
Then come the automated cuts.
Auphonic can detect silence, filler words, coughs and other respiratory noises, as well as longer foreground-music segments. Those detections can be applied automatically, converted to silence or exported as cut lists for another editing application. The Auphonic Editor allows users to move, disable or add cut regions rather than blindly accepting what the software decided.
That review function is essential.
A cut that is technically possible is not necessarily editorially desirable.
What it cannot decide
Auphonic can identify a pause.
It cannot know whether the pause carries meaning.
It can identify an “um” or “uh.” It cannot always know whether cutting that syllable will make a guest sound more polished or unnaturally abrupt.
It can classify sound as noise. It cannot know that the room tone, crowd sound or music underneath a conversation is something the producer deliberately wants the listener to experience.
And none of that is the hardest part of editing a reported podcast.
Auphonic cannot decide that an answer belongs 12 minutes earlier. It cannot determine that a guest contradicted a claim made elsewhere in the interview. It cannot recognize that a technically clean statement creates a legal or fairness problem. It cannot decide whether removing a section makes the episode more coherent while also making the guest’s position less accurate.
Those are editorial decisions.
Auphonic’s video features do not change that distinction. Automatic video cutting can remove corresponding video when audio sections are cut, and the service can now generate and burn captions into video. But it is not deciding which reporting document belongs on screen, which camera angle carries a moment, where B-roll improves comprehension or whether a visual edit misrepresents what happened.
It can automate production.
It cannot assume editorial responsibility.
Studio Voice deserves particular scrutiny in journalism
Auphonic’s newer Studio Voice feature goes beyond conventional noise reduction.
The company describes it as a system that reconstructs a clearer voice from the recording, including attempting to repair clipping, compression artifacts, distorted speech and missing high frequencies. Auphonic says both Studio Voice and its Voice AutoEQ rely on AI models that regenerate aspects of a voice.
For a casual podcast, that may be another cleanup option.
For journalism, generative or reconstructive audio deserves a stricter standard.
A reporter should be able to distinguish between removing hum from what a source actually said and algorithmically rebuilding portions of the sound of the source’s voice. If TRP uses Studio Voice on reported interviews, the processed output should be compared closely with the original before publication.
Cleaner is useful.
Faithful is mandatory.
The privacy question is more consequential than the waveform
Auphonic’s privacy policy says users retain rights to their uploaded content and that the company will not share that content or its associated metadata with third parties without permission, subject to the other circumstances described in the policy. It says normal productions and associated media are deleted after 21 days for audio and seven days for video and API productions.
But another section deserves attention from newsrooms.
Auphonic says users agree that uploaded content may be viewed or listened to by an employee or a person contracted to work on the company’s audio-processing algorithms. It also says selected audio or video files, or excerpts, may be retained internally to improve those algorithms.
That does not mean Auphonic is doing anything improper.
It means the newsroom has to match the tool to the material.
A public podcast episode that is going online tomorrow is one risk profile. An embargoed interview, confidential-source recording, protected minor, whistleblower conversation or unpublished investigative evidence is another.
Auphonic also says paying customers can request a Data Processing Agreement. Before using the service for sensitive journalism, I would want a clearer answer about whether newsroom customers can opt out of human review and internal retention for algorithm development.
That question matters more than whether the automated leveler saves me 15 minutes.
What it costs
Auphonic currently gives registered users two free processing hours per month. Those hours do not roll over, and free productions include an Auphonic jingle. Its recurring plans currently offer 9, 21, 45, 100 and 250 processing hours per month, with larger-volume options available.
One-time credits do not expire.
That matters for an independent producer whose workload varies. A recurring subscription works differently: unused recurring hours reset rather than accumulating.
Free users can test multitrack processing, but the free multitrack limit is shorter than 20 minutes.
Podcasters should check the current pricing page rather than treating any dollar figure in a review as permanent.
Where it fits in my workflow
For The Redemption Project, Auphonic currently makes the most sense as a finishing desk, not the entire control room.
The substantive edit still belongs in the part of the workflow where I make story decisions: deciding what stays, what goes, what moves, what requires context, where graphics and documents belong and whether an episode fairly represents the conversation.
After that, Auphonic could potentially handle a substantial portion of the technical pass: balance voices, reduce noise, hit the loudness target, prepare transcripts and captions, add metadata and generate final outputs.
If repeated tests prove reliable, presets and automated workflows could make that process much faster.
But the preset has to earn my trust.
A first production should not.
The next test needs recordings that challenge the system differently: clean studio audio, a weak remote guest, separate speaker tracks, a single mixed track, music, background noise, intentional silence and specialized names or legal terminology.
Then I want to listen.
Not just look at the waveform.
Auphonic appears capable of removing a meaningful amount of mechanical work from podcast production. It may also consolidate jobs that otherwise require several separate tools.
That still leaves the most important part untouched.
Automation can make the episode cleaner. A human still has to decide whether the episode is right.
Behind the Reporting: What One Auphonic Test Can — and Cannot — Establish
This first run establishes that the service works as an integrated production environment and that it can visibly alter leveling, generate a transcript and expose automated processing decisions for review.
It does not establish an error rate.
The first-use material does not tell us how often filler-word cuts are correct, how accurately names are transcribed, how much audible improvement listeners perceive, whether Studio Voice changes vocal character in edge cases or how the system performs across recording environments.
That is why this is labeled a first look.
The next testing standard
The useful test is not “does Auphonic work?”
It clearly processes files.
The useful test is whether the automated decisions remain dependable when the source material changes. TRP’s next runs should deliberately include clean studio speech, poor remote speech, crosstalk, independent speaker tracks, a single mixed track, intentional pauses, music and video.
Each run should preserve the original file so the same passages can be compared before and after processing.
For automatic cuts, the question should be particularly strict:
Would I have made this edit myself if the algorithm had not suggested it?
If not, turn it back on.
Promotional relationship
TRP initiated contact with Auphonic seeking the opportunity to test the product in a real podcast workflow.
Auphonic supplied promotional processing credits. In direct correspondence, the company later clarified that the credits were one-time credits that do not expire.
Auphonic also offered access to its affiliate program, which pays a 20% commission on qualifying sales.
No affiliate link is used in this review.
That relationship is disclosed because a review is more useful when readers know not only what the company provided, but what financial relationship it offered.
Sources
Auphonic Documentation — current platform overview
Auphonic — Singletrack Post Production Algorithms
Auphonic — Multitrack Post Production Algorithms
Auphonic — current pricing and credit structure
Auphonic — Studio Voice Beta, July 21, 2026
Auphonic — Burn-In Subtitles, July 29, 2026
Auphonic — Automatic Video Cutting, April 15, 2026
TRP original material: Brandon L. Burley’s Sept. 1 first-use Auphonic screenshots and production review.
Direct company correspondence: Auphonic Support/Georg Holzmann correspondence with TRP, Aug. 29-31, 2026, concerning promotional credits, expiration and affiliate-program offer.
I am a retired detective and criminal justice / government educator based in Tennessee. I founded The Redemption Project, as a place to focus on civics, rebuild non-partisan trust, and provide educational and emotional grace while learning about the news. I also have a column in Knox TN Today. My reporting and commentary have also appeared in other outlets including; Governing, The Arizona Capitol Times, South Florida Sun Sentinel, Police1, among other state and regional outlets.
