You can find the test job at the bottom.
THE FILES YOU'LL HAVE ACCESS TO ARE CONFIDENTIAL AND THEY'RE NOT TO BE SHARED! BREAKING THIS RULE WILL LEAD TO THE WORKER'S REMOVAL FROM THE GT TEAM.
What do we do?
We provide audio transcription services, which means that we convert audio and video files into text.
In our system, you can pick the files you prefer to work on. You must have a PayPal
account as we do payments only with PayPal
. Payments are done every weekend, on Friday
. Also we cover PayPal fee so you will receive exactly amount which you earned and there will be no fee to receive payment. To request payment go to WORK area to STATS page. There you can request for payment.
- Every time a client uploads a file, you will receive an e-mail. After that, you need to go to the transcription jobs page and choose the files you want to transcribe.
- There you will also see the information about the turnaround time, text format (clean or full verbatim) and timestamping type, if one is required.
- After your transcription is uploaded, you will see your current earnings on the stats page.
- Payments are done every weekend, on Friday. There will also be a statistics page where you will see how much you have earned for each file.
- All of your transcriptions are going to be checked by editors. You will receive a rating and comments to help you improve your work.
- If your average rating is better than 4.5, you can try to become an editor.
- Sometimes, the quality of audio or video files is horrible, so please let us know about that. We will contact the client and provide a refund for that file.
- Transcribe everything in US spelling unless the client asks for a different spelling in his comments.
Text format descriptions
The text is transcribed exactly as it sounds and includes all the utterances of the speakers.
- Speech errors. "I went to the bank on Tu-Thursday."
- False starts. "I, um, wanted- I wanted to be a musician."
- Filler words such as "like", "you know", "yeah", "mm hmm", "uh huh", "umm", "uh, um", "sort of", "so", "oh" et cetera.
- Slang words, such as gonna, wanna, gotta, kinda, 'cause, should be kept in the transcription. In addition, "you know", "I mean" should also be transcribed.
- Stutters and repetitions.
Only use these forms for the affirmative/negative:
Mm-hmm (affirmative) or Mm-mm (negative)
Uh-huh (affirmative) or Uh-uh (negative)
The transcribed text does not include:
- Speech errors
- False starts (unless they add information)
- Filler words. Note that some of them can have a different function. For example, in the phrase "Oh my God", "oh" is not a filler word, but a part of the expression, therefore it must be kept.
- Slang words must be written as "going to" instead of "gonna", "want to" instead of "wanna", et cetera.
- "Yeah", "yep", "yap", "yup" must be written as "yes"; "alright" must be written as "all right".
- Never spell "Ok" or "OK". It must always be spelled as "Okay".
- Never change spoken contractions (e.g. do not → don't or don't → do not)
- Don't start phrases with conjunctions in clean verbatim. If you really need to add the conjunction, just expand the phrase. For example: I went outside, but forgot to bring my umbrella.
Note: omit all the "mm-hmm", "yeah", "yes" reactions to retain a fluent text, unless they are answers to given questions. [Mind the verbatim ex. Don't write "mm-hmm" as an answer in clean verbatim, replace it with "yes"] Also, "like" can be kept if it is not a filler word, for example: "I paid something like 200 bucks" or when it introduces a quote: I was like, "Okay".
DO NOT remove filler words if they change the meaning of the phrase. E.g. I was kind of happy; It was like three days ago- leave them in
Speaker 1: I went to the store yesterday.
Speaker 2: Yeah.
Speaker 1: I wanted to buy some bread.
Speaker 2: Mm-hmm.
Speaker 1: There was none left. Can you believe that?
Speaker 2: Yeah.
Speaker 1: I went to the store yesterday. I wanted to buy some bread. There was none left. Can you believe that?
Speaker 2: Yes.
Timestamping time format must be: [00:00:00] (always in bold)
ALWAYS CONSIDER THE WHOLE FILE WHEN YOU'RE TIME STAMPING.
For example, if you do the 20-30 minutes part, your time stamping should start at 20, not 00
Every 2 minutes:
Every time the speaker changes:
NOTE: If there's a comment next to the audio file saying, "Please use the embedded time" or "burned-in time", you will need to download the file in order to watch the video and use the correct time.
1) In general, short and logical, grammatically correct sentences are preferred. When a speaker is using conjunctions like 'and', 'so', or 'but' to connect longer stretches of thought, it's often a good idea to create sentence divisions in those places. Also, don't forget to cut out the conjunctions in those places when they're not necessary.
2) Longer speeches should be separated into smaller paragraphs. Paragraphs shouldn't be longer than 500 symbols (about 100 words or 3-4 lines in the transcription tool).
3) Never summarize the speech in the audio you are transcribing.
4) Do not correct errors made by the speakers. Use [sic] instead. For example: "All of my childrens [sic] are in school".
EXCEPTION: Correct egregious phonetic and pronunciation errors that inhibit readability or understanding
Example: if a speaker pronounces "refrigerator, washer and dryer" as "refrigurator, washar and dryear", please use the correct word and spelling based on your context of the audio.
NOTE: Editors should not penalize transcribers for using/not using [sic]
5) Everything with a time-stamp, (e.g. [unintelligible 00:04:24]; [inaudible 00:02:24]) should be bolded.
6) Nothing else besides the speaker labels and marks with time-stamps should be bolded. So [laughs], [coughs], [laughter] or [crosstalk] shouldn't be bolded.
7) Never use () instead of 
8) Every sentence should end with a punctuation mark.
9) The beginning of every sentence should be capitalized.
10) Use '--' instead of '...'. Example: "I wanted to say something, but…" becomes "I wanted to say something, but--"
11) Expansions for "it's", "that's", "you've", et cetera are not considered mistakes, but it's preferable to transcribe them exactly as they are spoken. Example: If the speaker says "it's", it should be kept as "it's" regardless of the verbatim
12) Never use exclamation marks.
13) Speaker labels
- If there is more than one speaker, their names (if they are known) must be used. If not - mark them as "Speaker 1", "Speaker 2" and so on.
- Always use a speaker label, even if there's only one speaker. Speaker:
- Make each speaker's role in the audio as descriptive as possible. For example, "Interviewer", "Interviewee", "Host", "Reviewer", "Caller", "Receiver", et cetera.
- If the speaker's full name is used at some point (for example, David Butterfield:), later in the transcription, the last name can be dropped (for example, David Butterfield: becomes just David:). The same goes for the titles (Doctor, Pastor, et cetera). This is optional and is not considered to be an error.
- If you cannot identify who exactly is speaking, add a question mark before the speaker label. For example, ?Speaker 2, ?Interviewee 3. But if you cannot identify who is speaking throughout most of the transcript, it's better to then mark the speakers with a generic label like Interviewee or Participant. Still, it's important that you are able to at least identify the person asking questions or leading the discussion.
- Speaker labels must be written in bold, followed by a colon and a space. NEVER use the Tab button, only one click of the Space button.
Speaker 1: Some text.
Speaker 2: Some more text.
NOTE: Occasionally customers dictate instructions to format the transcription while they are speaking. These instructions should be followed when possible, but never transcribed.
Follow customer requests for spoken directions such as new paragraph, comma, period or bullet point (use a dash). Do not type out the instruction.
14) Italicize film, book, magazine, song titles, as well as artworks, plays, TV and radio programs, foreign expressions et cetera. Example: I watched an episode of Friends the other day.
NOTE: There is no need to italicize social media sites, company names, the Bible, books within the Bible, versions of it or other sacred writings (Koran, New Testament, Genesis, et cetera). If a speaker spells a word or a name, do this: My name is Anna. A-N-N-A.
NOTE: Abbreviations and acronyms should not contain dashes. Right: USA. Wrong: U-S-A
NOTE: Always research the proper capitalization e.g. iPhone, UCLA, SaaS
- Percentages should be written as "%". Example: 2%, 50%. But if the speaker says "A small percentage of the responders", never replace "percentage" by "%"
- Always write links like this: www.facebook.com/clientpage Never write them like this: w w w dot facebook dot com slash client page
- Use international monetary symbols to mark currency (in most cases). Do this: $50, €155. Standard abbreviations (USD, EUR) are also acceptable.
- DO NOT spell out amounts. Right: $195 Wrong: a hundred and 95 dollars
- Proper nouns: Article III of the Constitution, Genesis 1:1
- Measurement: eight inches, 10' x 4', 223 pounds, 6'4"
15) Sound events
- Sound events that are significant to the audio should also be noted. Use brackets [ ] for notes. The notes are always written in lower case regardless of the position in a sentence.
- For example: [laughter] - when more than one person laughs, separate line; [laughs] - when one speaker laughs, written in the same line; [screaming], [background noise] - separate line; [applause], [silence] - separate line; [crying], [sarcastically], [crosstalk] – same line (when 2 or more speakers all talk at the same time) et cetera should be used appropriately.
If a foreign language or a word (in this case, a language that is not English) is spoken, mark it as [foreign language], [foreign word] or [French language], [German language], et cetera if it can be identified.
Note: Authors/titles/personalities/common sense things are an exception from this rule and must always be researched. For example: carpe diem, Jacques Chirac, Frédéric Chopin, alea jacta est et cetera.
- [pause 00:00:00] is used to demonstrate a pause significant in a speech. It must be longer than 10 seconds for it to be marked.
- [phone rings] – when the phone rings once; [phone ringing] - when the phone keeps ringing; [dial tone] - when you hear the dial tone on the line - These tags should be on a separate line
- When the audio is cut or edited, use [sound cut] on a separate line
- Spell out numbers from 0-9: zero, one, two, three, four, five, six, seven, eight, nine (this does not include phone numbers).
Numbers from 10 and up are written in numerals: 10, 11, 20, 25, 99, et cetera.
NOTE: It is okay to break this rule if it improves readability or consistency (e.g. in a list of items)
- Thousands should be written in digits. Do this: 600,000, 5,000. Note that there's a difference between "a thousand" (spelled out) and "one thousand" (digits)
- Always spell out big numbers like "million", "billion", "trillion", et cetera. Do this: 40 million, 10 billion. Numbers under 10 should be spelled out (two million).
- Times of the day and dates: always capitalize AM and PM. Do this: 2:45 PM, 5:00 AM. When using o'clock, spell out the numbers: eleven o'clock, nine o'clock.
- Time periods should be written with an apostrophe: '60s, '70s style; however, "that man is in his 70s" (no apostrophe)
When the speaker uses symbols or equations (like a Math lecture) it's the editor's job to save transcription in Word format and to insert those formulas and symbols and then send that Word version to firstname.lastname@example.org. That's because our tool does not support symbols or formulas. Transcribers should do their best to transcribe those things using the tool, like writing formulas using regular symbols x = x + 2.
17) Double dashes or a single dash (mind the verbatim type)
Use double dashes --
When there is a change of thought (false start) in mid-sentence. Do this: Weren't we going to-- Okay, never mind.
Use a single dash -
When there is an interruption, but the speaker continues his thought. Do this:
E. g. 1:
Speaker 1: I thought he was-
Speaker 2: Yeah.
Speaker 1: - gone.
E. g. 2:
Speaker: Weren't we going to- heading to London?
When someone stutters. Do this: Why is this m-m-moist?
18) Quotation marks
- Double quotation marks "" are used whenever there is a direct quotation. Example: She said, "Don't bother telling me". Quotations are introduced by a comma, not a colon.
- Internal dialogues are seen as direct quotations as well. Example: And then I thought to myself, "What if I can't make it?" (Notice how the question mark is inside of the quotation. This is because the quotation itself is a question. However, if the rest of the sentence is a question containing a quotation, then the question mark would be outside of the quotation marks).
- Do not use quotation marks in indirect quotations (when the speaker paraphrases what has been said). Example: The man said that he was in a hurry.
19) If you cannot hear what word is being said, mark that as inaudible or unintelligible and specify the time.
- Use [inaudible 00:00:00] when it is impossible to hear what is being said due to interferences in the audio.
- Use [unintelligible 00:00:00] when you cannot understand what is being said due to the speaker's manner of speech, accent, et cetera.
20) Do NOT make up your own markings. Only use the ones that you find in the guidelines.
DO NOT use: [incomprehensible]; [indistinct]; [inunderstandable] et cetera
21) If there are curse words in the audio, transcribe them word for word
22) Onto vs. On to
Rule 1: In general, use onto as one word to mean "on top of", "to a position on", "upon"
Example: I climbed onto the roof.
Rule 2: When onto means "fully aware of", "informed about"
Example: I'm onto your plan.
Rule 3: Use on to, two words, when on is part of the verb.
Example: Please log on to our site.
23) If you do not prepare the transcriptions according to these requirements, you might be removed from the team of transcribers.
Ratings which are given by editors:
1 - from 0% to 80% accuracy
2 - from 80% to 85% accuracy
3 - from 85% to 90% accuracy
4 - from 90% to 95% accuracy
5 - from 95% to 100% accuracy
24) Always use et cetera unless the client asks otherwise in the comment section (not bolded).
Transcribers should know that mistakes like the following will be harshly penalized by editors. Along with accuracy, editors will be rating your files based on your grammar mistakes and/or lack of research.
this/these; their/there/they're; it's/its; lets/let's et cetera
Were going to see what there proposition was. (transcriber)
We're going to see what their proposition was. (edited version)
[unintelligible 00:05:12], the Russian rhythmic gymnast won the golden medal in Rio. (transcriber)
Yana Kudryavtseva, the Russian rhythmic gymnast won the golden medal in Rio. (edited version)
Congo's former president, [unintelligible 00:15:59] visited the Pentagon in 1983. (transcriber)
Congo's former president, Mobutu Sese Seko Kuku Ngbendu Wa Za Banga visited the Pentagon in 1983. (edited version)
- Recurring mistakes will also be harshly penalized!
- If a transcriber finishes 3 transcriptions and has 3.4 or a lower average rating, he/she will be removed from the team.
If you are interested in working with us, you need to read the provided instructions again and then finish our test job (clean verbatim; no timestamping needed), which you can download here:
Let us know if you have any questions.