Privacy
I keep this site light on data. You can browse every job without an account, and the site sets no cookies. This page lists everything the website, the API, and the MCP server (the AI Trainer Jobs app in ChatGPT and other AI apps) collect, why, who else receives it, how long it is kept, and how to get it deleted.
Effective and last updated November 4, 2026.
Who I am
AI Trainer Jobs (aitrainerjobs.co) is run by me, Danilo, a developer from Brazil. I build and run it on my own, so I decide how your data is used. For any privacy question or request, email danilo@aitrainerjobs.co.
Summary
- No account, no cookies, no ad trackers, and no profiles of visitors.
- I never store your IP address. My code uses it only in memory, to count requests per minute for rate limits.
- The MCP server and the API don't store what you search for. Their results contain only public job and company data.
- I don't sell or rent personal data, and I don't share it with advertisers.
The ChatGPT app and MCP server
The MCP server at /mcp is what ChatGPT, Claude, and other AI apps call when you use AI Trainer Jobs in a chat. It needs no account, no key, and no login. It only reads data: none of its tools can change anything or send anything for you.
I never receive your chat, your name, your email, or your AI account. The AI app decides when to call a tool and which values to send. I receive only those values. The table below lists every tool, what it accepts, and what it returns.
| Tool | Inputs (all optional unless noted) | What it returns | Inputs stored? |
|---|---|---|---|
search_jobs | query (search words, up to 100 characters), category, min_hourly_pay, pay_type, eligible_country (a two-letter country code), limit, offset | The number of matching jobs and up to 20 jobs. Each job has its id, title, company, category, pay, location, languages, level, hours, skills, short summary, posted date, job page link, and apply link. | No |
get_job | id (required, the job id from search_jobs) | The same job fields, plus whether the job is open, its responsibilities and requirements, and the company's job description. | No |
list_companies | None | Each listed company with its link, tagline, number of open jobs, typical pay, worker rating, and how it pays. | n/a |
get_company | name (required, a company name, up to 60 characters) | The same company fields, plus what it does, key facts, payment details, hiring steps, what workers say, common questions, sources, and the date the facts were checked. | No |
I use the inputs only to look up jobs or companies and answer the call. My code doesn't save them to a database and doesn't write them to logs. If a job id or company name doesn't exist, the error message repeats it back to the AI app, and that is all. If the AI app adds other details to a request, such as your language or approximate location, the server ignores them. The outputs are public data that anyone can see on this site. They never contain personal data about you.
To run the MCP server, these things also happen:
- Rate limits. I count requests per IP address for one minute, in memory only, and never save the address. AI apps send their tool calls from their own servers, so the address is usually the AI company's, not yours.
- Error logs. If the server fails to handle a request, it logs a short error message. It doesn't log the request itself.
- Cloudflare request logs. Cloudflare records each request to the site, as described under Cloudflare below. Tool inputs travel in the request body, not in the web address, and my code never logs the body.
- Links you open. Job links and apply links in the results point to this site. If you open one, it counts like any other visit or apply click, as described below.
The AI app's own privacy policy covers your chats and your account. For ChatGPT, that is the OpenAI privacy policy. For Claude, it is the Anthropic privacy policy.
The API
The API serves the same public job data with no account and no key. It works like the MCP server, with one difference. Its filters, such as the search words in q or a country code, are part of the web address. So they appear in Cloudflare's request logs, and Cloudflare caches the response for that address for 60 seconds. The cache holds the public results, not who asked. My code doesn't store the filters. Rate limits work the same way as for the MCP server.
What the website collects
- Page views. When you open a page, the site adds 1 to that day's total of page views. For the first page of a visit, it also adds 1 to that day's count for the site that sent you (its domain name only, such as google.com) and for your country. Your country comes from your connection. Only these daily totals are stored. There is no row per visitor, no IP address, and nothing that tells visitors apart.
- Apply clicks. Apply buttons go through a link on this site,
/go/, which sends you on to the company. For each click, I store the job, the time, your country, the address of the page you came from, and your browser's user agent (the browser and system name your browser sends). I don't store your IP address. Clicks from known bots are not stored. - Job alert signups. Email alerts are paused, and the signup form is hidden. If you signed up before, I store your email address, the categories you picked, how often you wanted emails, your country, and the dates you signed up or changed them. When the form is on, it uses Cloudflare Turnstile to check that you're not a bot. Turnstile loads only when you click into the form, and the check sends your IP address to Cloudflare.
- Analytics. Two analytics services count visits without cookies: Cloudflare Web Analytics and Simple Analytics. Their scripts load on every page and report the page, the site that sent you, and basic browser and device details to those services. Simple Analytics says it doesn't collect IP addresses. See analytics choices below.
The site has no other forms. The pay calculator and converters run on the numbers you type, which go into the page address and are not stored.
What stays in your browser
The site saves three small items in your browser's local storage. They are never sent to me, and they aren't cookies.
- The list of jobs you opened and marked as applied, with dates, so the list can show them. On a job page, you can undo an "applied" mark.
- The time of your last visit, so new jobs since then get a "New" label.
- Your light or dark theme, if you picked one.
They stay on your device until you clear this site's data in your browser.
What is public
The stats page shows totals only: visits and page views per day, apply clicks per day and per job, the countries visits come from, and the sites that link here. It never shows anything about one person.
Why I use this data
- Tool and API inputs: to find the jobs and companies you asked for.
- IP addresses: to apply rate limits and keep the service up for everyone.
- Page views and apply clicks: to see which pages and jobs people use, to spot bot traffic, and to show the public totals on the stats page.
- Signup data: only to send the job alerts you asked for, once alerts start.
- Error and request logs: to find and fix problems.
I don't use any of it for ads, and I don't combine it into profiles of people.
Who receives data
- Cloudflare hosts the site, the API, and the MCP server, and stores the data above in its database. Like any host, Cloudflare handles your IP address and request details to deliver pages and block attacks, and it records a log of each request (such as the address, time, and response). It also runs Web Analytics and Turnstile. Workers AI, Cloudflare's AI service, writes the job summaries from public job posts and never receives data about you. See the Cloudflare privacy policy and the Turnstile privacy addendum.
- Simple Analytics receives the page view details its script reports. See the Simple Analytics privacy policy.
- OpenAI and other AI apps send the tool calls to the MCP server and receive the results. They receive public job and company data from me, nothing about you.
- Companies you apply to. When you click Apply, your browser goes to the company's own site, such as micro1, Mercor, or Handshake. Some links add my referral code, which tells the company you came from here. The code identifies this site, not you. From that point, the company's privacy policy applies.
- Directory badges. The footer on most pages shows badges from sites that list AI Trainer Jobs. Your browser loads those images from the directory sites, so they receive your IP address and browser details.
No email service receives your address yet. Before alerts start, I'll name the email service here and update the date at the top.
How long I keep it
- MCP and API inputs: not kept. They exist only while the request runs.
- IP addresses for rate limits: used for a one-minute count in memory, never written to storage.
- Cloudflare request and error logs: kept by Cloudflare for at most 7 days, per Cloudflare's documentation.
- Apply click records: 12 months. The site deletes older records automatically every day.
- Daily page view and visit totals: 12 months. The site deletes older totals automatically every day.
- Signup data: until you ask me to delete it.
- Browser storage: until you clear it.
- Analytics services and directory sites keep data under their own policies.
Your choices and rights
You can ask to see, correct, or delete any data I hold about you, or ask me to stop using it. Email danilo@aitrainerjobs.co. For signup data, write from the address you signed up with. I'll reply within 30 days, usually much sooner. You won't be treated differently for asking.
- Job alerts. Email me to remove your address, and I'll delete it. Every alert email will have a one-click unsubscribe link when sending starts.
- Apply clicks and page views. These records have no name, email, or IP address, so I usually can't tell which ones are yours. If you tell me the job and the time you clicked, I'll delete the matching record.
- Browser storage. Clear this site's data in your browser settings.
- Analytics. A content blocker or turning off JavaScript stops both analytics scripts. Simple Analytics doesn't count your visits when your browser sends Do Not Track. The site's own daily page view totals don't check Do Not Track, because they are anonymous counts with nothing that tells visitors apart.
- MCP server. You can turn off or remove the AI Trainer Jobs app in your AI app's settings. Your chat history is managed by the AI app.
If you're not happy with my answer, you can also complain to the data protection authority where you live. In Brazil, that's the ANPD.
Children
This site is for adults looking for paid work. It isn't meant for children, and I don't knowingly collect personal data from anyone under 18. If you think a child gave me personal data, email me and I'll delete it.
Where the data is processed
I run the site from Brazil. Cloudflare serves it from data centers around the world, so your requests may be handled in another country, and the services above may process data in the United States and elsewhere. They do so under their own privacy policies and safeguards.
Security
The site, the API, and the MCP server use HTTPS. Only I have access to the Cloudflare account and the database. The best protection is collecting little, so I never store IP addresses, and the MCP server and API store nothing about the people who use them. No system is perfectly secure. If a breach ever affects your data, I'll tell you and the authorities the law requires.
Changes to this page
If I change how I handle data, I'll update this page and the date at the top. If the change affects your email address, I'll tell you by email first.