Transcription Jobs

Transcription is the process of converting audio (or video) into written text. You’ll listen to conversations, interviews, lectures, court proceedings, podcasts, medical dictations, or YouTube videos and type them up – either word-for-word (verbatim) or “clean read” (lightly edited to remove “ums” and repetitions).

Typical clients include podcasters, YouTubers, NGOs, universities, legal practices, healthcare providers, market researchers, media houses, and businesses that need accurate records of meetings, webinars, and calls. South African accents and multilingual content are in demand, so local transcribers who can navigate South African English, plus languages such as isiZulu, isiXhosa, Afrikaans, Sesotho, and Setswana, are becoming in valuable.

Yes – both as full-time employment at local language service providers, and as freelance/contract work via South African or international platforms. Examples of legitimate companies operating in or hiring from South Africa include Way With Words, a long-standing transcription/annotation firm with roots and operations touching Cape Town and broader Africa.

Red flags should include “jobs” that ask you to pay to apply or to “unlock” tests; offers with guaranteed high pay for very little work; clients pushing you to communicate outside official systems, and then asking for free “trial” projects; and vague companies with no website, no physical presence, and no client reviews. Rather, choose established platforms (global and local), look for clear pay structures, and read third-party reviews before making a commitment.

The main types of transcription jobs include:

• general transcription: podcasts, interviews, webinars, meetings;
• media/captioning/subtitling: timed text for video, often with style guides;
• legal transcriptions: hearings, depositions, court recordings – requires precision and confidentiality;
• medical transcriptions: clinical dictation – requires medical terminology, sometimes credentials;
• market research: focus groups, user interviews – often multi-speaker, with accents; and
• speech data / AI annotation projects: labelling audio snippets or verifying machine transcription –  often offered in high volume, and is fairly repetitive to complete.

For all of these job types, you’ll need:

• excellent spelling/grammar and attention to detail.;
• a good typing speed (~60–80+ WPM helps) and accuracy;
• accent comprehension (SA and international);
• familiarity with transcription software (foot pedals optional);
• ability to rapidly research names, terms, and brands;
• ability to follow client style guides and deadlines;
  and
• basic audio cleanup (e.g., using Audacity) is a bonus.

Rates vary widely. Many global marketplaces, such as Go Transcript, publish client-side pricing at around US$0.99 to US$1.50 per audio minute (what the customer pays), while freelancer pay is often a slice of that. For instance, pricing references for client-side rates commonly cluster in the region of US$0.99/minute for some services, and calculators explaining rate factors are publicly available; freelancer shares are lower depending on the platform’s cut and job type.

Your income could swing with all the following:

• your approval level (some platforms gate higher-paid work);
• hours you’re actually assigned (queues fluctuate);
• audio difficulty (clean audio pays less; complex accents/multi-speaker pays more);
  and
• niches (legal/medical can pay more with experience).

A realistic beginner range (working part-time) may equate to a few thousand rand per month; while a busy, skilled general transcriber working near full-time across multiple sources could push their earnings into mid-five figures in rand – especially if they specialise, add captioning/subtitling, or win direct clients. (Your mileage will vary with speed, accuracy, and job flow.)

The National Minimum Wage (NMW), effective 01 March 2025, increased to R28.79 per hour. Note: many transcription gigs classify you as an independent contractor, not an employee, so NMW may not apply in the same way as for permanent employment – but it’s still a useful benchmark when evaluating offers.

It’s crucial to combine global platforms (i.e. those that welcome South African freelancers – well-known marketplaces and language vendors with open tests), with SA-connected language companies like Way With Words (applications open periodically, and they’re strong on accuracy and SA accent familiarity).

Reach out to podcasts, agencies, researchers, law firms, and production houses, as well as LinkedIn (search under “transcriber,” “captioner,” “subtitler,” “QA transcription,” and “annotation”). Lastly, be sure to check out research/insight firms, as these may be running focus groups, locally.

Quality control measures generally consist of a short audio clip (two to 10 minutes), instructions or a style guide, and a deadline. You’ll be evaluated on formatting, speaker labelling, timestamps, punctuation, ideal spelling of names and brands, and the importance of following the instructions provided.

Many platforms use waitlists, however you can increase your odds of securing work by:

• taking every training opportunity or test seriously;
• scoring high marks for accuracy and speed;
• being available for unpopular time windows (overnight, and weekends);
  and
• building a niche (e.g., medical/legal, subtitles with precise timing, or a multilingual offering).

You must-have toolkit for transcription work includes:

• stable internet (preferably fibre);
• a quiet space and good over-ear headphones;
• transcription software (e.g., oTranscribe, Express Scribe, Descript, or browser-based editors used by each platform);
• a text expander or AutoCorrect for common words and phrases;
• a foot pedal (optional but useful for volume players);
• a basic audio cleanup tool (e.g., Audacity) to reduce noise when permitted;
• a password manager for client logins; and
• cloud backup for work files (and a clear deletion policy for privacy compliance. More on POPIA appears below).

Many companies pre-transcribe using automatic speech recognition (ASR). Your job therefore becomes human editing – i.e. improving accuracy, adding punctuation, fixing names, and any formatting. Learning to edit fast and accurately is key to this type of work.

Most online transcription gigs classify you as an independent contractor (self-employed). This means no PAYE deductions – and you are required to handle your own tax.

Independent contractors generally register and submit returns and, if they meet the thresholds, pay provisional tax twice a year (with a final annual return). SARS provides an official Guide To Provisional Tax outlining who is a provisional taxpayer, the IRP6 process, due dates, and calculations. You’ll also find practical freelancer-focused explainers (e.g., on provisional tax, allowable deductions, and record-keeping) that highlight typical differences from PAYE employees, and common pitfalls.

If you’re truly self-employed, typical deductions may include part of your home office costs (subject to SARS rules), internet, hardware, software, headsets, and professional services. Keep invoices and logs and, when unsure, consult a registered tax practitioner.

This is not generally required. Many freelancers operate as sole proprietors. However, you could opt for a (Pty) Ltd for reasons such as limited liability or branding. This is a personal, financial, and legal decision, so you’ll need to consult a professional to ensure you do it properly.

The Protection of Personal Information Act (POPIA) is South Africa’s data protection law. If you handle recordings with personal information, you must process it lawfully, securely, and for a valid purpose – and delete/retain that data appropriately. Summaries of POPIA spell out your duties (consent, security safeguards, breach notifications, and data subject rights). There are full resources, and the text of the Act available publicly to help you understand these obligations in detail.

Essentially, POPIA requires you to use secure storage (i.e. encrypted drives/cloud with access controls), to keep client data only as long as this is necessary (i.e. follow client retention policies), to avoid sharing raw files via insecure channels (i.e. to be extremely wary of public links), to anonymise where required (i.e. remove phone numbers/addresses if instructed), and to have a breach plan in place (i.e. notify clients promptly if something goes wrong).

Intro guides also compare POPIA to frameworks such as GDPR, and stress proportional safeguards suited to SA’s context – which are useful when you work with international clients who ask about compliance.

Common payment methods include PayPal, which allows you to withdraw to your South African bank account (fees/exchange rates apply); Payoneer, which offers global receiving accounts and local withdrawals; Wise

 (formerly TransferWise), that sees some clients paying you directly into your ZAR account, or into a Wise account – from which you can then make a withdrawal; and direct Swift transfers for larger corporate clients. Before making your selection, compare fees and FX rates, and keep a list of invoices and tidy records for SARS.

This will vary – it could be weekly, bi-weekly, or monthly, depending on the platform or client. Many well-known marketplaces may pay weekly for approved work, while others pay after a job is accepted. PS: Always read payout FAQs before starting, so these are clear to you from the outset.

Firstly, check whether you are required to create a “verbatim” or a “clean read” script. Verbatim includes fillers (“um,” “you know”), false starts, stutters, repetitions, and non-lexical sounds ([laughter], [sigh]), while a clean read needs such filler words removed, tidies grammar lightly, but still preserves meaning. Be sure to include regular timestamps (e.g., every 30 to 60 seconds) or speaker-change timestamps. Captioning/subtitling also needs time-coded segments, where reading speed and line-length constraints are noted.

In general platforms tend to measure quality using aspects such as accuracy (spelling, names), style compliance, formatting, correct speaker IDs, punctuation, and turnaround time. Repeated accuracy issues can limit your access to important and well-paying jobs.

You’ll log into a job board, claim files, download audio, open the platform editor (or your own), and work through the file. You’ll pause, rewind, slow down tricky sections, insert timestamps, and research names/terms as you go. Many editors let you playback with hotkeys so you can keep your hands on the keyboard.

A common range to transcribe an hour of clean audio is three to five hours; faster if you’re very experienced and slower for multi-speaker, accented, or noisy recordings. Editing ASR  may be faster – but still requires careful listening.

Specialisation is likely to improve your rate, as well as your acceptance for higher-tier queues. Notable niches that pay better than some others, include legal, medical, technical/financial, captioning/subtitling, and AI data/quality.

For general transcription, there are no mandatory qualifications. For legal and medical, some clients prefer or require relevant experience or specialised training. Courses can help you get familiar with style guides and expectations, but are not mandatory for most general work.

Consider: audio difficulty, turnaround time, speaker count, timestamps, verbatim vs clean, and subject matter. Many South African freelancers quote in ZAR per audio minute or per project, referencing public client-side pricing bands as market anchors (remember: platforms take a cut; direct clients pay for your full service and relationship).

Rate calculators and blog comparisons show how client pricing is often structured – which can be most useful for positioning. Also note, that you can add a “rush fee” of +25–50% for <24-hour delivery when a client demands work on a longer file, or on complex content, at short notice.

For direct clients, at a minimum, you’ll need to include the following in your paperwork:

• scope (verbatim/clean read), timestamps, format;
• turnaround time and delivery format;
• confidentiality/POPIA compliance and data retention/deletion;
• revisions policy (e.g., one round of minor edits);
  and
• payment terms (rate, currency, due date, late fees).

Platforms generally bundle these in their Terms and Conditions, while direct clients should sign your agreement or a mutual NDA.

Most companies run QA reviews. Your score affects access to higher-paying jobs. Expect feedback on style adherence, names, timestamps, and punctuation; and keep a personal “style bible” of each client’s particular quirks.

It certainly is, as it will expand the scope of the work you can take on. It generally involves timing text to frames, splitting lines properly, managing reading speeds (characters per second), and following broadcast or platform (e.g., Netflix/YouTube) style guides. The learning curve raises your value and can certainly improve your earnings. Listen in here for more

Beginners should avoid submitting work without checking spelling of names/brands; ignoring style guides (e.g., incorrect timestamp formats); over-editing when the brief calls for verbatim; poor audio hygiene (not using noise reduction when this is allowed); missing speakers in group conversations; accepting rushed deadlines that compromise the quality you are able to deliver; and not backing up (securely!) OR failing to delete files per policy (POPIA risk).

To improve your work within a month, you can seek feedback from another professional, read guides from different platforms, compile vocabulary lists and style guides for repeat clients, and build macros – such as hotkeys for timestamps, speaker labels, and common notes. It can also help quite a bit to shadow real files, by doing short, varied audios each day to practise.

Working from South Africa can be beneficial for four distinct reasons:

• accent agility: South Africans often handle a wide range of accents (local + international), which clients value highly;
• time zone (SAST, UTC+2): convenient overlap with Europe and partial overlap with the US for overnight turnarounds;
• cost of living vs USD/GBP earnings: foreign-currency income can be attractive – but watch exchange and withdrawal fees; and
• local compliance: POPIA awareness is a plus – overseas clients often ask for your data protection approach.

Aim to land your own direct clients by word of mouth (when you offer efficient turnaround and reliability), via partnerships with podcast editors, PR firms, video producers, by posting short tips on LinkedIn, about transcription tips and case studies, and by putting together a portfolio of non-confidential samples of your work. You can also display your niche area on LinkedIn, such as “I specialise in market research focus groups and finance webinars”, as just one example.

When pitching for work, create copy similar to the below for best effect:

Hi [Name].

I help [industry] teams turn audio/video into publication-ready text with [verbatim/clean read] accuracy, consistent timestamps, and style-guide compliance.

Recent projects include [type], typically delivered within [TAT]. For [file type/length], my rate is [ZAR/USD per audio minute], which includes [QA pass/one revision].

I’m POPIA-aware (secure storage, minimal retention) and comfortable signing NDAs.

May I transcribe a short sample to match your tone and format?

Your sincerely

It’s a good idea to draw up a structure for each project, that confirms scope, clarifies turnaround time and file format, and includes a preparatory stage where you establish details about the client and their audio file. Next up perform your transcription, proofread the document, and then deliver it securely – together with an accompanying invoice. Remember, as a final stage, to delete/archive per POPIA and client policy.

It is a good idea to reference public client-side pricing pages and calculators to understand what clients pay, and then to set your independent contractor rates based on your value (specialist knowledge, turnaround, and formatting complexity).

Market overviews and calculators help you explain why a complex, multi-speaker, 24-hour turnaround costs more than clean one-on-ones with a 3-day deadline.

Scale up by offering captioning/subtitling as an add-on, by moving into QA/Editor roles (where you review others’ transcripts), by building a micro-team (vet and train two to three transcribers; you handle QA, client comms), by adding audio cleanup or light podcast editing as a service, and by positioning yourself as a research support partner (transcription + summary + pull quotes) or, alternatively, as a content ops partner for podcasts and webinars.

When you encounter unclear audio – use better headphones, slow playback, try high-quality noise reduction (if allowed), and insert [inaudible 00:03:19] or [crosstalk] sparingly when it’s truly unintelligible. Offer clients optional audio enhancement (at a fee), where appropriate.

When you encounter inconsistent style guides – keep a “client bible.” If something’s ambiguous, follow the example document they have provided. Consistency beats perfection.

What about a feast-or-famine workload? Diversify across two to three platforms and one to two direct clients. Save an emergency fund, and use a simple CRM or spreadsheet to forecast work and cashflow.

When you encounter payment delays (direct clients) – implement deposits (i.e. before you begin) on larger orders, milestone billing, and clear due dates (e.g., Net 7/14). For platform work, check out their payout schedules before you start.

Read more about the typical challenges experienced by transcribers, here so you can prepare in advance to take these on with aplomb.

While you should vet every opportunity, these pointers can help you begin your research:

Be sure to know your basic duties around lawful processing, security safeguards, retention/deletion, and data subject rights. Summaries of POPIA and the full text are publicly accessible. Next up, read up on IRP6 filings, estimates, due dates, and annual return alignment; and be sure to keep tidy tax records for SARS. Lastly, the National Minimum Wage was pegged at R28.79 per hour from 1 March 2025. This is a useful as a benchmark – even if you contract independently.

It certainly is, if you treat it like a craft and a micro-business. The most successful South African transcribers combine three things: accuracy and speed (anchored by solid tools and tight workflows); professionalism (POPIA-aware handling, reliable delivery, good communication); and smart positioning (niche skills –captioning, legal/medical, or industry familiarity, and a mix of platform and direct work). Start with good habits, build your niche, keep learning, and your earnings and client base are likely to grown incrementally with you.