In short
Open Job settings, then AI, and switch AI on. It needs the OpenAI integration from step 5.
Pick a model on the AI Models tab (GPT-5 Mini is the default). The Resume Prompt tab holds the instructions the AI follows.
Switch AI on
Choose a model
The AI Models tab shows five OpenAI models as cards. The model is the AI that writes your resume. The one with the tick is selected, and your choice is saved at once. You can also move between the cards with the arrow keys.
| Model | Best for |
|---|---|
| GPT-5 | The most capable. Best quality, highest price. |
| GPT-5 Mini Default | Fast and affordable. A good balance of quality, speed and cost. |
| GPT-5 Nano | Fastest and cheapest. Good for simple jobs and for keeping costs minimal. |
| GPT-4.1 | Precise and reliable. |
| GPT-4o Mini | Light and quick. |
Only models your OpenAI key can use are listed. If TimelyDo cannot reach OpenAI it lists all five and says so; open the section again to retry. A model can only be changed while AI is on.
What each model costs
Every card shows the price per 1 million tokens, for example $0.25 in, $2.00 out. Tokens are small pieces of text: about four characters, or three quarters of a word. In is what is sent to the model, out is what it writes. A thin bar on each card shows what a typical resume costs next to the most expensive model.
Below the cards a panel shows the chosen model in full:
| Price | Meaning |
|---|---|
| Input | What is sent: the prompt, your default resume and the job. |
| Cached input | The start of a prompt OpenAI has seen a moment ago is charged at this lower price. TimelyDo keeps the start of every request the same on purpose so this applies. |
| Output | What the model writes, including the thinking it does first. |
It also says a typical resume costs about a dollar amount, measured on a real run of about 3,400 tokens in and 3,400 out, and names the cheapest model. Press Compare all models for a table with input, cached, output and cost per resume for all five.
| Model | Input | Output | A typical resume |
|---|---|---|---|
| GPT-5 | $1.25 | $10.00 | about $0.038 |
| GPT-5 Mini | $0.25 | $2.00 | about $0.0076 |
| GPT-5 Nano | $0.05 | $0.40 | about $0.0015 |
| GPT-4.1 | $2.00 | $8.00 | about $0.034 |
| GPT-4o Mini | $0.15 | $0.60 | about $0.0026 |
Which model should I pick?
- Start with GPT-5 Mini. It is the default for a reason.
- Choose GPT-5 Nano if you generate many resumes and cost matters more than polish.
- Choose GPT-5 or GPT-4.1 for roles where the wording matters most, and you accept a higher price per resume.
- If a run says the answer was cut off or the resume and job together are too long, try a larger model.
The resume prompt
The Resume Prompt tab shows the instructions the AI follows, in a text box headed Resume prompt sent to the AI. The default is written by TimelyDo. In short, it tells the AI to:
- decide whether the job was posted by the company itself or by a recruiter, and use the company details only in the first case;
- tailor the headline, summary, experience and skills to the job, without inventing employers, dates, education or achievements;
- keep wording natural, concise and free of buzzwords;
- score how well your original resume matched the job and how well the new one does, from 0 to 100;
- return the new resume in exactly the same structure as yours.
You can edit it, up to 20,000 characters. A counter under the box shows the length.
- Save prompt saves your version. It applies to the next resume you generate.
- Reset to default appears once you have your own version, and brings the default back.
TimelyDo adds your default resume, the job and the company details after the prompt automatically, so you do not need placeholders in it.
What is and is not sent
Only when you press Generate resume with AI are these sent to OpenAI with your key: the prompt, your default resume without its look settings (colours, fonts, spacing), and the details of that one job and its company. Nothing else about you is sent.
Frequently asked questions
Does turning AI on cost money?
No. Nothing is sent to OpenAI until you press Generate resume with AI under a job.
Which model is cheapest?
GPT-5 Nano, at $0.05 per million input tokens and $0.40 per million output tokens, about $0.0015 for a typical resume. See the comparison table in the AI settings.
Which model gives the best resumes?
GPT-5 is the most capable and GPT-4.1 is precise. GPT-5 Mini is the default and works well for most jobs at about a fifth of the price of GPT-5.
What is a token and what is a million tokens?
A token is a small piece of text, roughly four characters. Prices are quoted per million tokens. A resume uses a few thousand, so it costs a small fraction of the per million price.
What is cached input?
When the start of a request is identical to one OpenAI saw a moment ago, that part is charged at a lower price. TimelyDo keeps the start of the request the same for every job, so later resumes can cost a little less.
Can I change the prompt?
Yes, on the Resume Prompt tab. Save it, and use Reset to default to go back to the original.
I cannot switch AI on
AI needs the OpenAI integration with a key OpenAI accepts. Connect it first, then come back to the AI section.
My model is missing from the list
Only the five offered models your OpenAI key can use are listed. If your key cannot use one, it is not shown. If OpenAI could not be reached, all five show with a note.
Was this article helpful?
Let us know so we can keep improving our documentation.