$176,910top of the range in New Jersey · middle $60,170 / yr
AI is transforming this role
Translators in the United States earn a median of $60,170 a year. Pay starts near $37,070. Pay reaches $176,910 at the top of the range in New Jersey, the best-paying state for this work among those with at least 500 people in the job.
Source: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Interpreters and Translators, SOC 27-3091). Last checked 9 September 2026.
Entry level
$37,070
Top of the range · New Jersey
$176,910
Education
Bachelor's degree + language fluency
Wages — U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Interpreters and Translators). Top of the range is the highest state-level figure among states with at least 500 people in the job. AI-impact rating is PayCrunch's editorial assessment. Updated September 2026.
🆕 New & Trending AI Tools for TranslatorReviewed September 2026
We track new AI-tool launches every week and refresh this list — here’s what’s gaining traction for Translator work right now.
NotebookLMNEWFree / $7.99 mo
Google tool that answers questions grounded only in the documents you give it — with citations.
How a Translator uses it: load a glossary, style guide, or source packet and keep terminology questions on those documents
GammaNEWFree / $9 mo
Generates polished slide decks and one-pagers from a prompt.
How a Translator uses it: turn a glossary or briefing into a one-pager for the next assignment
ChatGPTFree / $20 mo
AI assistant for glossaries, domain prep, and first-pass drafts you still edit.
How a Translator uses it: draft a glossary, prepare domain terms, or produce a first-pass draft you still fully post-edit — never a final unreviewed delivery
ClaudeFree / $20 mo
AI assistant known for careful writing and long-document analysis.
How a Translator uses it: compare a long source and draft for tone, omissions, and consistency before you deliver
PerplexityFree / $20 mo
AI search engine that answers with live, cited web sources.
How a Translator uses it: check a current term, institution name, or cited fact with sources before you lock it
Canva AIFree / $13 mo
Design platform with AI text-to-image, writing, and one-click layouts.
How a Translator uses it: make a simple client one-pager or rate sheet without a designer
Google GeminiFree / $20 mo
Google's AI assistant, built into Gmail, Docs, and Search.
How a Translator uses it: draft and reply inside Google Workspace and research without leaving the page
Microsoft CopilotFree / $30 mo
AI built into Word, Excel, PowerPoint, Outlook, and Teams.
How a Translator uses it: write documents, build spreadsheets, and summarize meetings inside Office
Words on a page
A translator takes a document in one language and writes it in another so a specific reader can use it. The morning might be a contract, a medical record, a technical manual, a film script, a software string, or a letter a family needs for a school. You read the source until you know what it is trying to do. You write the target in the register that reader expects. You check names, numbers, and terms against the glossary you and the client agreed on. Then you deliver. The scene is a desk, a file, and a deadline. Spoken interpreting, the live voice in a meeting or a booth, is a neighboring craft. It can share a talent for languages. It is a different way of working, and it is not the job this description follows.
Good translation is invisible and accountable at the same time. The reader should not stumble, and the reviewer should be able to see why you chose a word. You keep a question list for the client: a date that contradicts itself, a term the source uses two ways, a sentence that cannot mean what it seems to mean. You do not smooth those problems into pretty prose and hope nobody notices. You also do not show off. A legal filing does not want a literary flourish. A literary novel does not want the voice of a filing. Matching the document is the skill.
Specialties change the risk. Legal work can move money and liability, so you flag ambiguity instead of picking a convenient reading. Medical records and discharge papers have to stay faithful for a patient and a clinician, which means you translate what is written and you do not add a treatment plan of your own. Technical manuals have to match the product the engineer built. Marketing has to sell without inventing a claim the source never made. Software strings have to fit a button and still make sense. Literature has to sound like a person wrote it. Most working translators pick two of these and get genuinely good, rather than advertising every subject on earth.
The week is project-shaped. An agency sends a file with a due date and a style sheet. A direct client sends a messier folder and a story about why it matters. You estimate whether you can finish well, you say no when you cannot, and you protect the due date when you say yes. Revision is normal. A second translator or an editor marks the draft. You accept the marks that make the text righter, and you explain the marks that would make it wrong. That conversation is professional, not personal. People who need every sentence praised burn out, or they burn their clients.
A voluntary credential, and a separate court path
Certification from the American Translators Association is voluntary. The association grants it to translators who have met its requirements in a language pair. It proves that a national professional body recognizes your competence in that pair. It does not prove you can handle every subject, and it does not replace a portfolio. Many excellent translators work without it, especially in language pairs or niches where clients hire on samples and references. Many agencies still like to see it, because it shortens the argument about whether you are serious.
People prepare by translating, a lot, under conditions that resemble paid work: real deadlines, real revision, and a mentor or a peer who will mark the draft. A degree in translation, linguistics, or a subject you translate can help. So can years inside the subject itself, the way a nurse who later translates medical records already knows the ward. The credential comes after that practice, when you decide a voluntary certificate will help the clients you want. Ask translators in your language pair whether their clients ask for it. If they do, plan for it. If they hire on samples, invest in the samples first.
Court interpreters may be credentialed by a state court. That credential belongs to spoken work in a courtroom, and a state court is the body that grants it. It is a different path from the association's voluntary certification for written translation. Some people hold both, because they do both jobs. If your aim is the written desk, do not assume a court credential replaces a translation portfolio, and do not assume the reverse when a court is hiring an interpreter. Name the craft in the credential's own words. Clients and courts both notice when you blur them.
What either credential proves is trust from an outside body. What it cannot prove is taste, speed on a live deadline, or honesty when the source is unclear. Keep a short set of samples in your strongest pairs, with confidential material removed or replaced. Be ready to do a paid sample rather than a free novel. Be ready to say which subjects you will not touch. That boundary is more persuasive than a list of every industry you have ever glanced at.
How the work shows up
Three doors are common. An agency keeps a roster and sends you files. A company hires you in house, as the linguist for a product, a hospital, a law firm, or a government office. A direct practice means clients know your name and send work without a middle desk. Early on, the agency door is the one that opens. The downside is the rate and the lack of control. The upside is volume, a revision habit, and a way to learn which subjects you want. In-house work trades variety for a salary and a team. Direct work takes longer to build and, when it is real, lets you choose.
Hiring, when there is a job rather than a project, looks at language pairs, subject knowledge, and a sample. Put the pairs in the first line, with the direction you actually work. "Into English from Spanish and Portuguese" is clear. A vague "fluent in several languages" is not. Describe tools you use for translation memory and terminology, without turning the resume into a software catalog. Then give the subjects: clinical trials paperwork, patents, employee handbooks, novels. Offer a sample that matches the posting. A beautiful poem will not win a contract-translation job.
Freelance hiring is a slower conversation. You write to agencies that use your pair, you take a test assignment they assign, and you deliver it as if the client were already angry about the deadline. After that, reliability is the whole reputation. Answer, deliver, and warn early if a file is bigger than you thought. In-house interviews add a team question in disguise: can you explain a choice to a product manager who does not read the source? Practice that explanation. It is the same muscle as a translator's note, spoken across a table.
A warning on the neighboring craft: if the posting is mostly depositions, hospitals at the bedside, or a conference booth, it is interpreting, even if the word translator slipped into the title. You can still apply if you do that work. Do not apply by pretending a written portfolio is the same as live practice. The wage series blends the two. Your application should not.
Pay inside a combined language series
May 2025 Occupational Employment and Wage Statistics report these wages under Interpreters and Translators. The series covers spoken interpreting and written translation together, so the figures are the published picture for that combined occupation, used here with the written desk in mind. The range opens at $37,070. The middle of the national figures is $60,170. The step from the opening figure to that middle is $23,100. The high end of the published range is $176,910, which sits $116,740 above the national middle. That high end belongs to a state that is not in the median list. Read the medians first, in order.
Maryland's median wage is $88,550. New York's median wage is $84,090. Massachusetts's median wage is $72,750. Colorado's median wage is $71,790. California's median wage is $70,770. Maryland's median is the highest median, and it sits $28,380 above the national middle of $60,170. New York follows in this order. Massachusetts, Colorado, and California are next, each a median, each above the national middle. Maryland is first. California is last. None of these medians is the high end of the range.
New Jersey's high end, Maryland's median
New Jersey holds the high end of the published range at $176,910. Maryland holds the highest median at $88,550. New Jersey is absent from the median list, so there is no New Jersey middle wage to quote from these figures. The high end and the medians are different statistics. $176,910 is the top of the published range. $88,550 is a middle wage.
The spread between the highest published state median and the lowest published state median is $43,240. That gap is large next to a national middle of $60,170, which is why a translator comparing notes with a friend in another state can think one of them has the wage wrong. Often both are looking at medians. The $116,740 climb from the national middle to the New Jersey high end is a different fact, about the top of a range, and it should not be used as a nickname for Maryland's $88,550.
Setting a year against the figures
Freelance translators often think in projects. The Bureau figures are annual wages. Translate the conversation back to the year before you decide you are underpaid or triumphant. A year that lands near $37,070 is the opening published level: early clients, a narrow pair, a lot of revision you do not get paid extra to perform. The $23,100 step to $60,170 is the move toward the national middle. It usually follows a repeat client, a subject you are known for, and the ability to decline work that pays too little for the risk. Ask an agency what a steady year looks like for someone at your level, then lay that picture beside $37,070 and $60,170.
In-house offers are easier to compare because they already look like wages. Put the offer beside $60,170 and beside the state median, in this order: Maryland $88,550, New York $84,090, Massachusetts $72,750, Colorado $71,790, California $70,770. If the desk is in Maryland, the highest median is $88,550, which is $28,380 above the national middle. You can say you are anchoring to that median, not to New Jersey's $176,910. If the desk is in California, anchor to $70,770, and do not import Maryland's median just because it is higher. Direct clients in a state with no median in this list should be compared with the national middle of $60,170 unless the employer brings a local figure and explains it.
Use $176,910 rarely. It is the high end of the published range in New Jersey, $116,740 above the national middle. It may describe a scarce language pair, a privileged in-house post, or a practice that has become a small firm. It is a foolish anchor for a new freelancer's rate card. If you are already earning near a high state median and a client wants you exclusively, you can mention the distance between $60,170 and $176,910 as the shape of the published range, then ask what exclusivity is worth inside that shape. Do not demand the high end because a chart exists. Demand a reason that matches the work.
Commission-style agency splits and rush premiums are easy places to invent numbers. Do not. If a contract states a split or a rush amount, use the contract's own words and still check the likely year against $37,070, $60,170, and the relevant median. The $43,240 gap between the highest and lowest published state medians is a caution against comparing a friend's annual story with yours when you do not even share a state. Write the state, or write "national," before you write the dollar.
A longer practice
The early years are about becoming the person an agency trusts with a messy file and a real due date. You learn one or two subjects deeply enough that project managers stop explaining the basics. You learn to write a translator's note that saves a client from a bad assumption. Later, some translators stay freelance and raise the kind of work they accept until a year near $60,170, or near a state median, is normal. Some go in house and trade the feast-and-famine calendar for a salary and colleagues. Some become the reviewer who marks other people's drafts. That last role is still written work. It is a step up in responsibility, not a change into the booth.
A few people add interpreting because a client asks and because they are actually good at it. If you do, keep the labels clean. Written delivery stays under the translator name. Live spoken work is interpreting, and a courtroom may want its own credential from a state court. The voluntary association credential remains a credential for translation. The wage chart still blends them under one series, which is why you have to be the person who unblends them in a negotiation. Say which craft the offer is buying.
Keep the numbers in the same order you learned them. The range opens at $37,070. The national middle is $60,170. Maryland's $88,550 is the highest median, $28,380 above that middle, followed by New York, Massachusetts, Colorado, and California's $70,770. New Jersey's $176,910 is the high end of the published range, a different statistic, $116,740 above the national middle. Certification from the association is voluntary. A state court may credential an interpreter, which is a separate path. The career itself is a long attention to other people's sentences, rewritten so the right reader can act. Pay follows the readers who come back, and the boundaries you were willing to say out loud.
The top of Translator pay — and how to get there with AI
$176,910what Translator pay reaches in New Jersey
Highest state-level top-of-range annual wage for Interpreters and Translators, among states with at least 500 people in the job. U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025.
And the role it leads to — English Language and Literature Teachers, Postsecondary — reaches $216,190 in California.
$37,070entry$60,170middle$176,910top end
Rates at the top of this occupation go to translators who own a subject terminology, decide which material may never touch an outside system, and teach both to everyone else on the project.
Machine drafting has taken the easy half of this work and left the hard half untouched. Compiling terminology for legal or medical material, checking that technical terms stay consistent across revision after revision, resolving genuine conflicts about what a word or a practice means, judging content and context against the intended audience — none of that is drafted for you. What has changed is who does the first pass and how fast. A translator who builds a governed terminology bank, sets the rule for which documents go near a public model under the confidentiality code they work to, and trains colleagues on both becomes the person a client asks for by name rather than by rate.
Your playbook, by where you are now
Just startingBuild a bank, not a habit
Choose one domain and start a structured terminology file in Microsoft Excel or Microsoft Access with source, target, context and the authority behind each entry.
Record every conflict you resolve and how you resolved it, because that reasoning is what a reviewer will later challenge.
Compare your own rendering against a Claude or Gemini draft on the same passage, and write down where each went wrong and why.
Learn the file formats clients actually send, including HTML and XHTML markup, so tags never eat your delivery time.
Read your professional body's confidentiality code closely and decide, in writing, which categories of material you will never paste into an outside system.
What proves it: A domain terminology bank with sourced entries and a written confidentiality rule you apply consistently.
Realistic span: the first two years
A few years inBecome the consistency check
Offer terminology review on other people's work, which is the paid role that sits above straightforward drafting.
Build a check that flags where a technical term drifted between revisions, and run it before delivery rather than after complaint.
Take the certification your subject area recognises, since legal and medical clients screen on it before they read a sample.
Use Grammarly and a model for mechanical passes only, and keep the meaning decisions in your own hands and on record.
Quote separately for terminology work, review and post-editing instead of folding everything into one rate.
What proves it: Paid terminology or review work on a project you did not translate yourself.
Realistic span: years three through seven
ExperiencedSet the standard others follow
Write the style and terminology guide a client's whole vendor pool has to follow, and charge for maintaining it.
Run the training when an agency adopts a drafting assistant, covering what to check, what to reject and what never to send.
Take on the material where an inaccurate rendering carries consequence, since consequence is what raises a rate.
Look at New Jersey, the strongest paying state for this occupation, and at teaching translation at postsecondary level as a second income line.
Move to direct clients where you set terms, keeping agency work only as filler between projects.
What proves it: A client style and terminology standard in force, plus a training role when tools change.
Realistic span: year eight and beyond
The next 90 days
Take the next ninety days and turn the terminology living in your head into a file somebody else could use. Pick one subject area. Every time you settle a term, log the source, the target, the context it applies to, the authority you relied on, and the alternative you rejected. Then run a small experiment: give a model the same passage you have already translated and note precisely which terms it got wrong. That list is evidence, and evidence is what lets you tell an agency what post-editing on their material genuinely costs. Translators who can quantify where machine drafts fail in their domain are the ones who get asked to set the rules rather than work under them.
Wage figures: BLS OEWS, May 2025. The playbook is PayCrunch editorial guidance, not a guarantee of pay or placement.
Every figure is the national median from the U.S. Bureau of Labor Statistics (OEWS) shown on that role’s own page.
Never used AI before? Start here (2 minutes).
Open a CAT tool with a built-in MT engine, starting with Trados Studio or the free Smartcat. Load a real document, turn on the DeepL or GPT plugin, and post-edit the machine draft segment by segment. You are not typing from scratch anymore; you are correcting, and your job is to catch what the machine got wrong.
For skills and terminology, use free tools deliberately: ChatGPT or Claude to explain an unfamiliar domain term in context, DeepL's free tier to compare renderings, and Linguee or IATE for verified terminology. Keep anything confidential out of free tools and reserve those for public or practice text.
The one rule, forever: Never paste confidential, NDA-covered, or personal client data into free public MT engines or chatbots, which often retain and train on inputs. Use tools with contractual no-retention terms (DeepL Pro, enterprise CAT plugins) or on-premise engines. And never trust MT fluency: it invents plausible-sounding errors, so verify every segment against the source. In legal and medical work a mistranslation can cause real harm.
The plays — exact steps, exact prompts
Do these in order. Each one is copy-paste ready. You do not need to know anything about AI going in.
1
Post-edit machine translation at 2-3x your old speed
Why this pays: You're paid per word or per hour; MTPE lets you deliver far more words per day at the same quality, and post-editing rates times your volume beat old from-scratch rates. Speed on the mechanical 80% is what frees time for the premium work.
RWS Trados StudioDeepL ProSmartcat
1
Set up Trados Studio (or free Smartcat) with a DeepL Pro or GPT plugin so each segment pre-fills with a machine draft you edit rather than type.
2
Build a repeatable post-editing pass so you catch MT's classic failure modes.
Copy-paste this prompt
You are a senior [legal] translator reviewing a machine translation from [source language] to [target language]. For the text below, flag any: mistranslated terms of art, added or omitted content, wrong register or formality, false cognates, and number or date errors. List each issue with the segment and a corrected rendering. Do not silently rewrite fluent-but-wrong text.
Source:
[paste source]
MT draft:
[paste MT]
Use on non-confidential text or with a no-retention enterprise tool. This catches the 'fluent lie' MT is famous for.
3
Track your words-per-hour before and after. Renegotiate to per-word MTPE rates once you can prove your throughput.
What you'll haveTwo to three times the daily word count at professional quality, the volume that lifts annual income toward $176,910.
2
Own a premium specialty MT can't touch
Why this pays: Generic translation rates fall as MT improves; specialized legal, medical, patent, and financial translation holds premium per-word rates because errors are costly and liability is real. Specialists sit at the top of the pay band.
ChatGPTClaudeIATE / Termium
1
Pick one high-stakes domain (contracts, clinical trials, patents, financial filings) and use Claude or ChatGPT to build a fast onboarding curriculum in its terminology and conventions.
2
Have AI teach you the domain's landmines in your language pair.
Copy-paste this prompt
Act as a translation trainer for [English to Spanish] [medical device] translation. Give me: (1) 30 must-know terms of art with the standard target-language equivalent and a common wrong translation to avoid, (2) the register and regulatory conventions (e.g., EU MDR instructions-for-use style), and (3) 5 example sentences where a literal MT rendering would be dangerously wrong, with the correct version.
Verify every term against an authoritative glossary (IATE, MedDRA, client TM) before using it in paid work.
3
Advertise the specialty and certification (e.g., ATA certification in your pair). Charge specialist rates, not commodity rates.
What you'll haveA defensible niche where your rate rises instead of falling, the core of a top-of-range freelance income.
3
Transcreation: sell judgment, not word count
Why this pays: Transcreation (adapting marketing and creative copy culturally) is billed by the hour or project, not per word, at rates far above standard translation, because MT cannot make brand-and-culture calls. It's the highest-margin work a translator can take.
ChatGPTDeepL WriteClaude
1
Offer transcreation for ad copy, slogans, UX strings, and brand voice. Use ChatGPT or Claude to generate several culturally distinct options fast, then apply your human judgment to pick and refine.
2
Generate options, then choose and justify. The justification is what clients pay for.
Copy-paste this prompt
I'm transcreating this [English] marketing tagline for a [Mexican] audience: '[tagline]'. The brand voice is [playful, premium]. Give me 6 target-language options ranging from literal to fully reimagined, and for each note the connotation, cultural risk, and reading level. Then flag any that carry unintended slang or regional meaning.
AI proposes; you decide. Cultural and legal sign-off is your value and your responsibility.
3
Package it as a project rate with a rationale document. Clients pay for the strategy write-up, not the word count.
What you'll haveHourly or project rates several times your per-word rate on the creative work MT can't do.
4
Build a terminology and translation-memory asset that compounds
Why this pays: A clean translation memory (TM) and termbase mean higher match rates on every future job, so you get paid to translate text you've effectively already done. It's a moat that raises your effective hourly rate over time.
memoQApSIC XbenchChatGPT
1
In memoQ or Trados, maintain a per-client TM and termbase. Run ApSIC Xbench to catch terminology and consistency errors before delivery.
2
Use AI to mine a termbase from a client's existing bilingual material.
Copy-paste this prompt
Extract a bilingual glossary from these aligned [English/German] sentences. Output a table: source term | target term | part of speech | usage note. Prioritize domain-specific terms, product names, and anything a general dictionary would get wrong. Flag inconsistencies where the same source term is translated two different ways.
Human-review every entry; a wrong termbase entry propagates into every future job.
3
Reuse and grow the asset on every project. Your leverage grows as your TM does more of the work.
What you'll haveA rising effective hourly rate as your TM does more of the work, the compounding edge of an established specialist.
5
Add subtitling, transcription, and AI-dubbing revenue
Why this pays: Audiovisual localization is a growing, higher-rate segment. AI transcription and machine subtitling give you a fast first pass; you edit for timing, reading speed, and nuance, adding a second income stream without a second full skill-build.
OpenAI WhisperSubtitle EditElevenLabs
1
Use Whisper (via Subtitle Edit) to auto-transcribe and time source audio, then edit the subtitle file for reading speed, line breaks, and meaning.
2
Translate and condense subtitles to spec. AI helps you hit the character limits.
Copy-paste this prompt
Condense this subtitle line to a maximum of [42] characters per line, [2] lines, while preserving meaning and natural [target-language] phrasing, for a reading speed of about [17] characters per second. Original: '[line]'. Give 3 options at different compression levels.
Subtitling is constrained rewriting; verify timing and on-screen readability yourself.
3
Offer AI-dubbing review with ElevenLabs: let clients generate voice, and sell your linguistic QA and script adaptation on top.
What you'll haveA second revenue stream in audiovisual work, added days billed per month without leaving your language pair.
Your 12-month sequence to the top of the range
How the plays above stack into a path from median pay toward the $176,910 tier.
Month 1
Set up a CAT tool with an MT plugin and start post-editing real jobs. Measure your words-per-hour.
Months 2-3
Choose one premium specialty and build its termbase; begin studying toward certification in your pair.
Months 3-6
Add transcreation or subtitling as a higher-rate service; raise rates on repeat clients using your throughput data.
Months 6-12
Grow per-client TMs, earn certification, and shift your mix toward specialty and project-rate work over commodity per-word jobs.
Gear for this job
As an Amazon Associate, PayCrunch earns from qualifying purchases. Links to books and tools are for the job on this page; we only recommend what we’d use in the work.
Same live Jossey-Bass 3rd already on high-school-teacher / middle-school-teacher / math-teacher / test-prep-instructor / substitute-teacher / science-teacher / music-teacher / drama-teacher / adult-education-teacher / corporate-trainer / instructional-designer / stem-teacher / pe-teacher / speech-teacher / curriculum-developer / education-consultant / college-professor / assistant-principal / financial-literacy-educator / school-principal / vice-principal / homeschool-consultant / school-administrator / edtech-specialist / education-administrator / distance-learning-coordinator / capitol-police-officer / tsa-agent / piano-tuner / birth-doula (ASIN 1119712610). This leftover page is BLS Interpreters and Translators (SOC 27-3091); title is Own the Glossary, Then Teach It; play 1 is Post-edit machine translation at 2-3x your old speed; start-here is Open a CAT tool with a built-in MT engine; Month 1 is Set up a CAT tool with an MT plugin and start post-editing real jobs. Classroom technique for leftover glossary / teach-the-domain / specialty-instruction load — not leftover Wong as the lead (that is nanny / children-s-librarian) and not leftover Praxis as a dump. Confirm 1119712610. Live page HTTP 200, no PC_GEAR / amazon.com/dp / tag=paycrunch-20 at 2026-09-18 4:20 AM PT. Source page: pe-teacher.
What Translators earn by state
These are the Bureau of Labor Statistics’ own figures for Interpreters and Translators, state by state — not a cost-of-living adjustment applied to the national number. Only states employing at least 500 people in the occupation are shown, because a state median drawn from a handful of workers is noise rather than a signal.
Maryland
$88,550
highest of them · +47% vs the national median
Michigan
$45,310
lowest of the 26 states that qualify · -25% vs the national median
The same job pays $43,240 more a year at the median in Maryland than in Michigan — 95% higher. That gap is what the Bureau measured, before any question of what it costs to live in either place. The top-of-range figure quoted at the head of this page, $176,910, is a different statistic in a different place: it is the 90th-percentile wage in New Jersey. The state that pays the typical worker most and the state where the best-paid go highest are not always the same one.
Source: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025, SOC 27-3091. 26 states clear the 500-employee reporting floor for this occupation; those below it are left out rather than shown with a wide error band.
Free data. Use any of it.
PayCrunch publishes verified, BLS-sourced salary + AI-playbook data on 1,000+ professions — free, no signup.
It has already replaced the easy, low-stakes end: generic, high-volume text is increasingly raw machine translation. But demand for human translators in high-liability and creative work is holding, because MT still hallucinates, flattens tone, and can't be held accountable. The job is shifting from typing translations to editing machine output and owning the judgment calls. Translators who post-edit well and specialize are busier; those competing with the machine on generic text are being squeezed.
Is machine-translation post-editing 'real' translation, or a pay cut?
It's the mainstream of the industry now, and it's only a pay cut if you let it be. Priced right (a per-word MTPE rate times your higher throughput, or an hourly rate), it can pay more than old from-scratch work. The trap is accepting rock-bottom per-word MTPE rates on volume you can't actually sustain at quality.
Can I safely use ChatGPT or DeepL for client work?
Not for confidential or NDA-covered material on free tiers, which may retain and train on your input. Use paid or enterprise tiers with no-retention terms, or on-premise engines, for client data. Free tools are fine for public text, practice, and terminology research.
Which should I learn first, a CAT tool or the AI chatbots?
The CAT tool. It's where you'll actually work: it manages segments, TM, termbase, QA, and the MT plugin in one place. Trados Studio is the market standard; Smartcat and memoQ are strong alternatives. Layer ChatGPT and Claude on top for terminology, domain learning, and transcreation options.
How do I keep my rate up as MT gets better?
Move up the value chain. Specialize where errors are expensive (legal, medical, patent, financial), add transcreation and audiovisual work billed by project or hour, and get certified. Sell judgment, cultural nuance, and accountability, the things a client can't get from a machine, and use MT to do more of the mechanical work faster.
Methodology & sources
Salary (median, 10th, top of the range) — U.S. Bureau of Labor Statistics, OEWS.
By state — the Bureau of Labor Statistics’ own state medians, limited to states employing at least 500 people in the occupation. No cost-of-living arithmetic is applied to a wage anywhere on this page.
The plays — PayCrunch's own step-by-step guidance using publicly available AI tools. Tool names/URLs are real and current as of August 2026; prompts written to work as-is. Verify any professional output before relying on it.