How to start a podcast: setup, recording, editing, publishing

Most abandoned podcasts did not fail at the microphone. They failed because nobody wrote down what the show was, so every episode became a fresh argument with yourself about what it should be.

Answer four things in plain sentences. Who is this for. What do they get out of an episode. How long is it. How often does it come out. Then commit to a format that survives a bad week: solo monologue, one recurring interview slot, or two hosts talking through a topic. Anything with heavy production per episode is a promise you will regret in month three.

Batch your first three episodes before you publish any of them. Apple and Spotify listeners who find you on day one want something to binge, and you get a buffer for the week life gets in the way.

One decent USB microphone, a pair of closed-back headphones, and a quiet room. That is the whole starting kit. Interfaces, mixers, and boom arms are upgrades, not requirements, and none of them fix a room that echoes.

Spend the money on the mic before anything else, then treat the room for free. Record in the smallest carpeted space you have, with soft things around you: a bed, a wardrobe of hanging clothes, a sofa in the corner. Hard parallel walls and glass are what make home audio sound like a home.

Wear the headphones while recording. Not for vanity, for catching the problem live: a fan, a fridge compressor, a phone buzzing on the desk. Fixing that in the room takes two seconds. Fixing it in editing takes an hour and never fully works.

Get within a hand's width of the mic and slightly off axis so your plosives go past it rather than into it. Proximity does more for perceived quality than any plugin.

Solo, you can record straight into whatever editor you plan to use. GarageBand is free on a Mac and perfectly capable of capturing a clean voice track. Audacity does the same job on Windows and Linux and costs nothing. Set levels so your loudest laugh peaks somewhere below the top of the meter, then leave the gain alone for the rest of the session.

Remote interviews are where shows get ruined. Do not record the call itself if you can avoid it, because you are recording a compressed stream that drops out whenever someone's wifi hiccups. Use a tool that records each person locally and uploads the clean file afterwards. Riverside and SquadCast both do this, and both give you separate tracks per speaker, which is the thing that makes editing an interview bearable.

Whatever you use, run a thirty second test with the guest before the real conversation, and record a local backup on your own machine as well. Guests do not do second takes.

Ask everyone to wear headphones. Without them, your voice comes out of their speakers and back into their mic, and no amount of processing removes that properly.

There are two honest paths here and picking the wrong one costs you months.

If you want to edit by reading, Descript turns the recording into a transcript and deleting a sentence deletes the audio. It is genuinely faster for interviews and rambling solo takes. The gotcha is that aggressive filler-word removal flattens how a person actually talks, so use it lightly on anything where the performance matters.

If you want a real editor, Reaper is a proper multitrack DAW for a one-time license and no subscription, and Hindenburg is built specifically for spoken word with automatic leveling that suits voice. Audacity remains the free answer that does everything a talk show needs, with a dated interface as the price.

If editing is the part you dread, Alitu automates the whole chain: cleanup, levels, music, export. You give up control and get episodes that actually ship.

Whatever you choose, keep your first edits boring. Cut the false starts, cut the tangent that goes nowhere, top and tail the episode, done. Do not learn compression and de-essing in week one. Auphonic will handle loudness, leveling, and noise reduction as a final pass better than a beginner will by hand.

Use your host or client's delivery specification. Apple recommends overall loudness around minus 16 LKFS, with a tolerance of 1 dB, and true peaks no higher than minus 1 dB FS. Common mono and stereo presets differ, so check the finished export instead of treating one target as a rule for every app.

You need a license for any music you put in a podcast, and popular songs are not available to you at any price you would pay. Use a library built for this. Epidemic Sound is the common choice for a subscription that covers podcast use while you are subscribed. Read what happens to episodes already published if you cancel, because that varies by library and it is the detail people miss.

Cover art is a hard requirement, not a nicety. Apple wants a square JPEG or PNG between 1400 and 3000 pixels on each side, RGB, and it needs to be legible as a thumbnail the size of a fingernail. Show name large, no tagline crammed along the bottom, no thin type.

A podcast is an RSS file at a URL. Apple, Spotify, Overcast, and everything else just read that file. Your host generates it and stores the audio, which is why the host decision is really a question about who controls your show.

Buzzsprout is the calmest first paid host for one show, billed on upload minutes with support that answers like a person. Transistor is the one to pick if you already know you will run more than one show, because unlimited shows sit on one bill. Both will issue a proper 301 redirect if you leave, which is the single most important thing to confirm before you pay anyone.

Spotify for Creators offers free hosting with optional monetization. You do not have to join its ad program to publish. It supports sponsor content and provides a 301 feed redirect if you leave; check eligibility and terms before relying on any platform income.

Once your first episode is live on the host, submit the feed yourself. Apple Podcasts Connect needs the RSS URL and an Apple ID, and review usually takes a few days. Spotify for Creators accepts the same feed and is normally live within hours. Do these in your own name and inbox, never a co-host's, never an agency's.

You do not need a website on day one, but you will want one. Podpage builds a site straight from the feed and updates it when you publish, which is the least work for the most useful result.

Publish on the same day and time every week. Consistency beats production quality for retention, and it is free.

Pull one or two short clips per episode for social. OpusClip does the cutting and captioning automatically from the long recording. Do not treat it as the whole strategy, and turn the auto-cleanup off for narrative shows.

Ignore download numbers for the first month. They will be small, the number is noisier than it looks, and reading them daily is a good way to quit. Look at whether people finish episodes and which topics bring anyone back.

Write useful show notes with accurate titles, links, and a clear summary. Add an accessible transcript when possible. Search engines can also index suitable video pages, so show notes are one part of discoverability rather than the only content a search engine can understand.