
To stitch podcast segments together, trim and level each segment as its own file, load them into the free Audio Joiner in running order, and export one episode. This online audio joiner decodes and joins everything in your browser, so your audio never leaves your device and there's no account to make. If all you need is a theme at the start and a sign-off at the end, the shorter guide on adding an intro and outro to audio covers that case.
A full episode might be a cold open, a theme, the host's welcome, an interview in two parts, an ad read or two, and an outro. Write that order down before you open any tool, so you catch problems early, like an ad read recorded on a different mic.
| # | Segment | Typical source | Prep notes |
|---|---|---|---|
| 01 | Cold open | A clip lifted from the interview | Cut to a clean sentence; leave a beat at the end |
| 02 | Intro theme | Mastered music file | Usually much louder than speech; normalize it |
| 03 | Host intro | Host mic, recorded separately | Trim edges, de-noise if needed |
| 04 | Pre-roll ad | Host-read ad | Level it to match the host intro |
| 05 | Interview, part 1 | Host and guest recordings | Shrink long pauses, keep natural pacing |
| 06 | Mid-roll ad | Host-read ad | Place it at a natural break, not mid-thought |
| 07 | Interview, part 2 | Host and guest recordings | Same prep as part 1 |
| 08 | Outro | Host sign-off, theme out | Fade the very end if the music needs it |
Ad slots follow a simple pattern: pre-roll before the main content, mid-roll at a natural break in the conversation, post-roll after it. Your show may use one, two or all three.
Name your files with a number up front (01-cold-open.wav, 02-theme.wav and so on) so they sort in order in your file picker. You can still reorder anything in the joiner with its arrow buttons.
An audio file joiner places files end to end. This one doesn't trim, clean up or add gaps, so each segment should be ready before it goes into the queue.
Because the joiner has no gap control, the space between segments is whatever you leave at their edges. When you trim, keep a short beat of room tone at the start and end of each segment. Room tone is the near-silence of the room you recorded in: the faint hiss and air that's present even when nobody is talking. A beat of it sounds like a natural pause, while a hard cut to digital silence can sound like a dropout. Start with roughly half a second and adjust by ear. If a segment was cut tight, you can also record a second or two of room tone as its own spacer file and drop it into the queue where a breath is needed.
A joined episode only sounds like one show if its parts sit at one loudness. The host mic, a guest's own recording, an ad read done on another day and a mastered music cue rarely arrive at the same volume. The joiner won't fix that. It never changes level, except for a peak limit it applies when a crossfade is on, so overlapping audio doesn't clip.
Before joining, run every segment through the Volume Normalizer in Loudness (LUFS) mode at the same target. Pick the preset you plan to publish at and use it for every file. With all segments at one loudness, the jumps at each seam come from the content, not the gain. How to normalize audio volume walks through the tool, and what is LUFS explains why perceived loudness, not peak level, is the number to match.
joined_ followed by a timestamp, with the extension of the format you picked.One slider value applies to every join in the queue. That matters for a podcast, because the seams don't all want the same treatment. Speech-to-speech seams, such as host intro into an ad read or interview part 1 into the mid-roll, sound best as plain cuts with room tone on either side. Crossfading two voices puts both on top of each other, and neither is easy to follow. Music seams, such as the theme flowing into the host's welcome, often sound better with an overlap.
The way around it is to build in sub-assemblies:
Exporting the in-between file as WAV keeps the extra pass from adding lossy encoding. Note that each overlap is capped at half the shorter of the two neighbouring clips.
Once the episode is one file, give it a finishing pass:
The segments weren't normalized to the same target, or one was skipped. Normalize the odd one out at the same preset as the rest and join again.
The trim was cut too tight. Go back to the Audio Cutter and re-cut that segment with a breath of room tone before the first word or after the last one. If a cut must stay where it is, a very short fade at that edge with Audio Fade, which works in 0.1-second steps, usually removes the click.
Each file is checked before it's decoded: up to 120 minutes or 2,000 MB on a desktop, and 45 minutes or 400 MB on phones and other low-memory devices. The joiner also has to hold every decoded segment and the joined result in memory at once. Splitting the episode into halves won't help, because the final join still produces the same long file. Very long episodes are best assembled on a desktop computer with more memory.
Find the break in the Audio Cutter and export the episode as two files, part 1 and part 2. Then join part 1, the ad read and part 2, with the crossfade Off. Normalize the ad read to the same target as the episode first.
The podcast workflow from recording to publishing puts this assembly step in context, from the first take to the upload. More guides like this one are in the how-to section.
How do I add an ad read in the middle of a podcast episode? Cut the finished episode into two files at a natural break using an audio cutter, then join part 1, the ad read and part 2 in that order with the crossfade off. Normalize the ad read to the same loudness target as the episode before joining so it doesn't jump out. A beat of room tone on either side of the ad makes the transition sound intentional rather than abrupt.
Does the audio joiner match the volume of my segments? No. The Vocal Cut Audio Joiner places files in sequence without changing their level. The only exception is a peak limit applied when a crossfade is on, so overlapping audio can't clip. To get consistent volume across a podcast, normalize every segment to the same LUFS target in a volume normalizer before joining, then normalize the finished episode once more.
Should a podcast episode be mono or stereo? A speech-only podcast is often fine in mono, since a single voice doesn't gain much from two channels. Stereo matters when the show uses music, sound design or panned voices. In the Vocal Cut Audio Joiner, the first file in the queue sets the channel count for the whole output, so put a stereo file first if you want a stereo episode.
How many segments can I join at once? The Vocal Cut Audio Joiner has no fixed limit on how many files you can queue; you need at least two. The practical limit is your device's memory, because every segment is decoded in the browser and held in memory along with the joined result. The total running time matters more than the number of files, so very long episodes work best on a desktop computer with plenty of memory.
Can I join MP3 and WAV segments together? Yes. You can mix formats in one queue, such as an MP3 ad read between WAV interview segments, and the browser decodes each one before joining. Everything is converted to the first file's sample rate and channel count, and the episode is exported as WAV, MP3 or Opus (Opus is offered only in browsers that can encode it, not currently Safari). Export as WAV first if you plan any further edits.