ITJobs Adds Tech Job Search and Applications to ChatGPT, Claude and Gemini

Looking for a tech job in Portugal just got shorter, at least for people who already start their day in a chatbot. ITJobs has launched an MCP server, a Model Context Protocol connection that lets people search listings, check how well their CV fits a role and submit a complete application without leaving ChatGPT, Claude or Gemini.

The interesting part isn't the search. Plenty of AI tools can already point you to job boards. Most then hand you off to an external web page the moment you want to apply, and that handoff is where the friction usually sits. ITJobs keeps that final step inside the conversation too.

MCP is worth a quick explanation if you haven't dealt with it yet. It's an open protocol that lets an assistant call outside services in real time. Instead of generating an answer purely from its training, the model can query a live source, act on what it finds and report back. For a job portal, that means the assistant reads current listings and writes to your account, so you aren't copying text between windows or retyping the same fields on every form. Fewer tabs, fewer repeated answers.

From listing to submitted CV in three messages

Setup is a one-time step: you link your ITJobs account to the assistant. After that, everything runs in plain language, closer to asking a recruiter about open roles than filling in a web form. ITJobs describes a typical sequence like this:

1. You ask what's new, for example remote roles with a visible salary. The assistant returns a list showing company, location and pay.

2. You pick one, such as a DevOps Engineer position, and ask the model to read the full requirements and draft a tailored cover message.

3. You review the draft and reply with a simple instruction to go ahead. The system submits the application with the CV stored on the platform and confirms delivery.

The third step matters most. According to ITJobs, nothing reaches a company without your explicit confirmation in the chat. That's the right default. An assistant drafting a message saves time; an assistant firing off applications without a checkpoint would be a problem. Here the review point is built into the flow rather than left to the user to remember.

Daily tracking and interview prep

The same connection covers the routine side of a search. The assistant can compare the CV on your account with the technical requirements of a specific role and point out where the two diverge. That's a quicker way to tell whether a posting is a stretch or a realistic match than reading it line by line, and it's most useful when you're applying to several roles that look similar on the surface.

Scheduled searches are part of the package too. The example ITJobs gives is a morning digest at 9:00 of new remote roles paying above €50,000. You can also ask for a side-by-side table of three listings covering technologies, salary and working arrangement. Beyond that, the assistant can review your application history, set follow-up reminders, pull information about companies and generate interview questions based on the wording of a posting. That last feature is probably the most practical for people preparing for technical rounds, since the questions come from what the employer actually wrote.

Carlos Andrade, Manager at ITJobs, ties the move to how people now begin their day. "We wanted that when someone asks about jobs, the answer comes from ITJobs, with real offers and the ability to apply right there," he says. He adds that "job searching has always been hard work" and that the goal was to bring the whole process into one simple conversation.

Availability and what to watch

The MCP server is available from today, 9 November 2026, free of charge for all registered users. It supports Claude, ChatGPT and Gemini, along with any other assistant that works with this open protocol.

My read is that the open standard matters more than the novelty. Because MCP isn't a proprietary plugin tied to one chatbot, the same connection should in principle keep working as assistants change, and you won't need a separate integration for each one. That's a sensible bet for a job portal, where you don't want to rebuild your workflow every time a new model appears.

A few caveats are worth keeping in mind. The assistant drafts from the CV stored on your account, so an outdated profile means weaker matches and weaker cover messages. Generated text still needs your edits, especially for roles where tone and specifics count. And concentrating your search inside a chatbot makes it easy to skim listings quickly, so read the full posting before you confirm an application.

The more useful thing to watch is whether other job platforms adopt the same protocol. If they do, the comparison will shift from which site has the most listings to which one an assistant can actually act on, and that's a different question for both job seekers and recruiters.