AI cover letter generator that stays honest

What the generator is allowed to see

The draft should be built from your resume and from one job description. It can rearrange a true detail so it sits next to the ad's phrase. It should not introduce a company, a degree, or a metric that is not already yours. If the output contains a sentence you would not say, delete it. That edit is free. If the whole draft is a costume, throw it out and write three sentences yourself. You do not owe the token a second try that makes the costume more confident.

Run Resume Score on the resume before you generate, so you are not paying to describe a PDF the parser could not read. After the letter exists, read it next to the resume once, checking dates and job titles. Then export both. The generator is not an ATS, and the letter is not Job Match Score. The match number, if you want it, is a separate check against the same posting, labeled as a match to that posting.

The letter does not repair a broken PDF

A cover letter is read by a person, often after the resume has already been stored as text. If Parse Rate on the resume is poor, the letter will not put the missing column back into the ATS. Fix the resume file first, with the free Resume Score. Then write the letter in the language of the ad.

The letter has one job: show you read this posting, using proof already on the resume. Two short stories beat a page of enthusiasm. Name the requirement, then the place you did the neighboring work, with the same dates the resume uses. If the dates disagree, the reader trusts neither document. Do not add an employer the resume does not contain. JobHoot's generator is held to that. A token, Rp 250, pays for a draft from the resume and the posting. It does not pay for a fictional career. Editing the draft is free. A new account starts with 50 tokens. The smallest purchase is 100 tokens, Rp 25.000.

Skip the letter when the employer said not to send one. Do not send an English template to a Bahasa Indonesia ad, or the reverse, unless the ad itself mixed languages and you are matching that mix for the tool names only.

After the draft appears

Read it aloud once. Any sentence you stumble on is a sentence to cut. Check every number against the resume. Check the company name against the ad, not against a previous letter you generated last week. The model will happily reuse a story that belonged to another posting if that story is what it saw. Your job is to notice. Delete the story that answers a requirement this ad did not ask.

Then export the letter and the resume as separate files with obvious names. Do not paste the letter into the resume's summary to "make one ATS document." They are different objects. The resume is what gets searched. The letter is what a person reads if they get that far. A high Resume Score does not require a letter. A letter does not raise Parse Rate. If someone sells you a generator that claims to do both with one click, they are selling a blur of the two names.

Read the nouns, then export two files

Read the draft aloud. Stumble, and you cut. Every number has to already be on the resume. The company name has to be this ad, not last week's. If a story answers a requirement this posting did not ask, delete the story. Export the letter and the resume as separate files with dull, obvious names. Do not paste the letter into the summary to invent a single ATS document. The resume is searched. The letter is read if a person gets that far.

A token pays for the draft, Rp 250, from material you already stored. It does not pay for a second career. Editing is free. Fifty tokens come with a new account. The guest Resume Score is still free and still comes first, because generating a letter about a PDF with no text layer describes a file the parser did not see. Job Match Score, if you want a number against this same posting, is a separate check with its own name. The generator is not that check.

If the draft names a tool you have not used, delete that clause before you export. The same rule applies to a metric, a city, and a job title. Two files leave your desk: the resume PDF that select-all can read, and the letter. You do not merge them to chase a single score.

The draft is a suggestion. The resume is the record. If they disagree, the resume wins and the letter changes.