Voice over is one of the better categories on Fiverr for a specific reason: the buyer can hear whether you’re good in eleven seconds. There’s no ambiguity, no taste argument, no “I’ll know it when I see it.” Your sample either sounds professional or it doesn’t. That’s brutal at the start and enormously useful once you’re competent, because it means you can win against sellers with more reviews on the strength of the audio alone.
It’s also a category where buyers reject deliveries for reasons that have nothing to do with your performance — the wrong file format, an audible room, a missed spec in the brief. Most of what follows is about eliminating those.
Table of Contents
Equipment: what’s actually good enough
The honest version, in three tiers. There is no tier zero. A phone or laptop mic will not pass on a paid job, and starting with one mostly teaches you that you can’t get orders.
Starting tier — roughly $100–$250 total. A decent USB condenser or dynamic microphone, a pop filter, and a stand. USB is fine here; the internet’s insistence that you need an audio interface on day one costs beginners money they should be spending on their room. Buy new or used from a known brand, plug it in, and get to work. Plenty of paid Fiverr voice over work is delivered on gear in this range and buyers cannot tell, because the room and the performance matter more than the capsule.
Closed-back headphones are not optional at any tier. You cannot hear your own mouth noise, plosives or room reflections on laptop speakers, and every one of those is a revision request waiting to happen. Budget $50–$100.
Working tier — roughly $400–$900 total. An XLR microphone plus a two-input audio interface. This is where most sellers doing voice over as real income land. The gain you get is not “better sound” in a way buyers consciously notice — it’s a lower noise floor, more headroom, and a signal that survives processing without falling apart. Practically, it means fewer takes and less repair work in post, which is a time saving, which is money.
Professional tier — $1,500 and up. A large-diaphragm condenser in the class buyers in commercial and broadcast work expect, a better preamp, proper acoustic treatment. Reach this tier when the work is paying for it, not before. Buying a $1,200 microphone into an untreated bedroom produces a beautifully detailed recording of a bedroom.
The rule underneath all three tiers: spend on your room before you upgrade your microphone. A $150 mic in a treated space outperforms a $1,000 mic in a bare one, every time, and it isn’t close.
Recording space on a budget
The enemy is reflection. Hard parallel surfaces — bare walls, windows, desks, floors — bounce your voice back into the mic milliseconds after the direct sound, and the result is that hollow, boxy quality buyers describe as “sounds like it was recorded at home.” Secondarily, the enemy is external noise: traffic, fridges, air conditioning, neighbours, plumbing.
What works, cheapest first:
A wardrobe full of clothes. Not a joke, and not a compromise. Hanging fabric is an excellent broadband absorber, and a full wardrobe is close to an anechoic box. Many working voice actors record in one. Cost: $0.
A duvet fort. Mic on a stand, heavy duvet draped over a frame behind and above you, more soft material on the floor. Ugly, effective, free.
Blankets and moving pads on the reflective surfaces. Identify the wall you’re facing, the wall behind you, and the desk surface. Cover them. Moving blankets are cheap and dense. $50–$100 gets a small space usable.
Acoustic foam and mineral wool panels. $150–$400 for a small room. Mineral wool panels substantially outperform thin foam tiles at low frequencies. Treat first reflection points and the corners.
A reflection filter around the mic. $50–$120. Does less than people hope, because it only addresses reflections arriving from behind the mic. Fine as a supplement, poor as a substitute for treating the room.
Things that don’t help: egg cartons, scattered foam applied for looks, and recording in a bathroom or kitchen.
Free noise control: turn off HVAC and the fridge for the take. Record when the street is quiet. Phone in another room. No jewellery or anything that rustles. Rug on hard floors.
Test your space honestly: record thirty seconds of silence, normalise it hard, and listen on headphones. Whatever you hear is what your buyer’s noise reduction will have to fight, and what you’ll be asked to fix.
Software
You need a recorder/editor and, eventually, a small chain of processing.
Free and genuinely sufficient to start: Audacity, or Reaper’s evaluation version (Reaper is inexpensive to license and is what a lot of professional voice actors actually use). Both record, edit and export everything Fiverr buyers ask for.
Paid, if and when you need it: Reaper’s full license, Adobe Audition, or Studio One. The reason to upgrade is workflow speed on long-form work — batch processing, better editing ergonomics, macros — not sound quality.
Processing. Keep it minimal and consistent: a high-pass filter to remove rumble below the voice, light de-essing if you’re sibilant, gentle compression, then normalisation to the target level. Aggressive processing is the most common beginner mistake — over-compressed, over-EQ’d, noise-reduced into a watery artefact-laden mess. A clean, lightly-treated raw recording beats a heavily-processed bad one every time.
Learn what loudness targets mean, because clients increasingly specify them. Broadcast and streaming clients may ask for a specific integrated loudness figure and a peak ceiling. If a brief specifies one and you don’t know how to hit it, you’ll deliver something that gets rejected for a reason you don’t understand. Most editors have a loudness meter built in; learn to read it.
Setting up a gig that converts
Narrow the service. “Professional voice over” competes with an enormous field. “Warm conversational male voice over for explainer videos” competes with far fewer, converts better, and can charge more, because the buyer recognises their own project in the description. Pick a lane based on what your voice genuinely does well — commercial, corporate narration, e-learning, audiobook, character and animation, IVR and phone systems, YouTube narration, documentary. You can add gigs later; you cannot be everything on day one and be chosen for anything.
The gig title’s first words matter most, because that’s what survives truncation in the search grid. Lead with the deliverable and the voice type.
Thumbnail. Voice over thumbnails have to work at grid size on a phone. A clean image of you at the mic, or clear legible text stating the voice type, both work. Avoid dense text and stock imagery of anonymous headsets.
Description. Write to a buyer deciding whether to spend money, not to a search engine. State plainly: what you sound like, what you deliver, in what formats, how many revisions, turnaround, whether commercial rights are included, and what you need from them (script, pronunciation guide, reference, deadline). Answer the buyer’s real question — “will this be right for my project, and will it be painless?”
Requirements field. Use it properly. Ask for the final script as text, the intended use, any pronunciation notes, a reference sample of the tone they want, and the deadline. Every question you ask here is a revision you don’t have later.
For how visibility works once the gig is live, see what impressions on Fiverr mean.
Samples and the demo reel
Your samples are your gig. Buyers will judge you on them almost exclusively.
Make them short. Fifteen to thirty seconds each. Buyers decide in the first few seconds; a ninety-second sample is a ninety-second sample nobody finishes.
Make several, one per use case, rather than one long medley. A buyer looking for e-learning narration should be able to hear e-learning narration without sitting through a car commercial first.
Use your own scripts. Do not record copyrighted ad scripts or read from published books to demo with. Write short original scripts in the style of the work you want, or use public-domain text. This matters legally and it also demonstrates you understand the format.
Record them at the quality you’ll deliver at. A polished demo followed by a rough delivery is the fastest way to a bad review.
Rerecord them as you improve. Most sellers’ samples are eighteen months behind their actual ability, which costs them orders every day.
Pricing and what to charge
Voice over pricing on Fiverr is typically structured per word or per finished minute, with the length band written into the package.
New, no reviews. Expect to start at the lower end of the category — often somewhere in the $5–$25 range for a short script — for your first handful of orders. Treat this as the cost of acquiring your first reviews, and set an explicit exit: five to ten orders, then raise. The mistake is not starting low; it’s staying there, because low prices attract the buyers who demand the most revisions per dollar.
After roughly 10–20 good reviews. Move to something like $25–$75 for a standard short script (up to a few hundred words), with extras priced separately. You should be turning down work at the old price by now.
Established, with a clear specialism and strong reviews. $75–$300+ per short project is achievable, and per-finished-minute pricing on long-form narration becomes the more sensible model. Audiobook and long-form e-learning work is usually quoted per finished hour and is a different business — high volume, long turnaround, and you should not take one on until your workflow is genuinely reliable.
Those bands are estimates based on typical listed prices in the category, not published data. Fiverr does not release per-category earnings figures.
Price your extras deliberately, because this is where the margin is:
- Commercial and broadcast usage rights. This is the big one. A voice over used in a paid advertising campaign is worth substantially more than the same recording used in an internal training video, and the industry prices usage separately from the recording as a matter of course. Define clearly what your base price covers — typically limited, non-broadcast use — and price broadcast, paid-media and perpetual-usage licences as extras. Sellers who bundle unlimited rights into a $20 gig are giving away the most valuable thing they own.
- Rush delivery.
- Extra revisions beyond the included count.
- Background music or a mixed final file.
- Sync to picture / timed to video.
- Multiple character voices.
- Source or raw unprocessed files.
Remember the arithmetic underneath all of this: Fiverr takes a flat 20% commission, on the base price, the extras and any tip. A $25 gig is $20 to you before withdrawal costs. Price accordingly.
Delivering files
Most avoidable disputes in this category are delivery-spec disputes, not performance disputes.
Formats. WAV is the standard deliverable for anything that will be edited further — 44.1kHz or 48kHz, 16-bit or 24-bit. 48kHz is the convention for anything going into video. MP3 at a high bitrate is fine for buyers who just want a finished file and are working on a phone or a limited connection. Delivering both, unless told otherwise, prevents a whole class of complaint and costs you nothing.
Ask which they want before you record, not after. Sample rate is the one thing you can’t cleanly fix in post without a quality loss, and a video editor who receives 44.1kHz for a 48kHz timeline will notice.
File naming. Name files descriptively — project, take, version. Buyers assembling a video from twelve separate line reads will remember whoever sent them final_final_2.wav and not fondly. Split multi-line scripts into separate files when the buyer is editing, and say so.
Revisions. State the included number in the gig and hold the line. Define what counts: a revision is a rerecord for a delivery issue or a direction change you were asked for. A script rewrite after delivery is new work and should be priced as new work. Say this in the gig description, politely, before it becomes an argument.
Raw and source files. Decide your policy and publish it. Many sellers deliver processed files as standard and offer raw unprocessed takes or the project session as a paid extra. Either policy is fine; an unstated policy is what causes trouble.
Deliver early where you can. On-time delivery is one of the performance signals Fiverr weighs, and buffer is free.
Why deliveries get rejected
In rough order of frequency:
- Audible room tone or background noise. Fix the space, not the file.
- Wrong format or sample rate. Ask upfront.
- Mispronounced names, brands, technical terms or acronyms. Ask for a pronunciation guide in your requirements. If the brief doesn’t include one, ask before recording — never guess a brand name.
- Wrong tone. The buyer wanted warm and conversational; you delivered polished announcer. Ask for a reference sample. This single question prevents most tone rejections.
- Pacing and read length. Buyers timing to video have a fixed duration. If they gave you one, hit it.
- Plosives, mouth clicks, sibilance. Pop filter, hydration, mic angle, and actually listening back on headphones before you send.
- Inconsistency across a long recording. Mic position and level drift between sessions. Mark your position; record long-form in as few sessions as you can.
- Missed instruction buried in the brief. Read the whole brief twice. Buyers hide the important line in paragraph four.
Scaling
Once you have reliable order flow, growth comes from four places.
Raise prices. The most direct lever, and the one sellers avoid longest. If you’re delivering on time with good reviews and a full queue, you are underpriced.
Sell longer-form work. One 45-minute e-learning module is worth many 100-word scripts and involves a fraction of the client-management overhead per dollar. Long-form is where per-hour economics improve dramatically.
Sell usage rights properly. Covered above, and it’s the highest-margin change available to a voice actor who currently bundles them.
Build repeat clients. A studio, agency or content producer who sends work monthly is worth more than any amount of search visibility. Deliver early, be easy to work with, and make it obvious you’re available for ongoing work.
Two things to be realistic about. Adding more gigs across many niches usually dilutes rather than multiplies — depth beats breadth in a category judged on sound. And synthetic voice tools have compressed the bottom of this market meaningfully; the cheap, generic, low-stakes read is the work most at risk. The defensible work is anything requiring direction, character, nuance, brand consistency or a human the client can talk to. Position there.
FAQ
What equipment do I need to start doing voice overs on Fiverr? A USB condenser or dynamic microphone, a pop filter, a stand and closed-back headphones — roughly $100–$250 for the mic setup plus $50–$100 for headphones. A phone or laptop microphone will not pass on paid work. Spend on treating your recording space before upgrading the microphone.
Do I need a soundproof booth? No. You need an absorbent space, which is a different thing. A wardrobe full of hanging clothes, a duvet fort, or blankets over the reflective surfaces around your mic will get you to a professional-sounding recording for nothing or close to it.
How much should I charge for voice over on Fiverr? Starting with no reviews, expect the lower end of the category for your first several orders, then raise once you have ten to twenty good reviews. Price commercial usage rights, rush delivery, extra revisions, sync-to-picture and source files as separate extras — that’s where the margin lives. (Specific ranges in the pricing section above are estimates, not published figures.)
What file formats should I deliver? WAV at 44.1kHz or 48kHz is the standard for anything the buyer will edit further; 48kHz is the convention for video. A high-bitrate MP3 covers buyers who just want a finished file. Delivering both unless told otherwise prevents a lot of complaints. Confirm the sample rate before you record.
Should I include commercial rights in my base price? Generally no. Usage is priced separately from the recording throughout the voice industry, and a voice over used in a paid ad campaign is worth far more than the same file used internally. State clearly what your base covers and sell broadcast, paid-media and perpetual licences as extras.
How much does Fiverr take from voice over orders? A flat 20% of the order value, including extras and tips. It’s the same at every seller level and it isn’t negotiable, so $25 gross is $20 to your balance before any withdrawal costs.
Why do buyers reject voice over deliveries? Most commonly audible room noise, wrong file format or sample rate, mispronounced names and brand terms, the wrong tone, and pacing that doesn’t fit the video. Almost all of these are prevented by asking better questions in your gig requirements field before you record.
How long are good voice over samples? Fifteen to thirty seconds each, and several of them — one per use case — rather than one long medley. Use original or public-domain scripts, never copyrighted ad copy or published books.
Can I do voice over on Fiverr with no experience? You can start with no client history, but not with no skill. The category is judged on audio, so buyers will hear the gap. Practise, build proper samples, and pick one specific niche before you launch.
VERIFY BEFORE PUBLISH
- Verified and safe: the flat 20% Fiverr seller commission applying to base price, extras and tips, at all seller levels, non-negotiable.
- All equipment, software, room treatment, file format, loudness and pricing-structure guidance is general voice-over industry practice, not Fiverr policy, and does not depend on platform mechanics. No Fiverr-specific verification needed.
- Price bands are explicitly labelled estimates. Fiverr publishes no per-category earnings or pricing data. Do not remove that framing or convert the estimates into stated facts.
- Not verified, deliberately omitted: current Fiverr Pro application criteria for voice over, any category-specific gig limits or per-level gig-count caps, and whether Fiverr currently applies any AI-voice disclosure or restriction policy in this category. If you want to cover AI-voice policy, check Fiverr’s Terms of Service and category guidelines first — it was not confirmed here and the article makes no claim about it.
- Fiverr Help Center pages returned HTTP 403 to automated fetching throughout this research round. Nothing on this page depends on those articles, but if you add anything about seller levels, Promoted Gigs or analytics to this piece, verify it manually first.
- The article references Fiverr’s requirements field, revisions setting and gig packages as existing features; confirm the current UI naming if you add screenshots.