# I Built an AI Agent That Found 2,002 Jobs For Me (No Code)

Source: https://qvantx.com/insights/i-built-an-ai-agent-that-found-2-002-jobs-for-me-no-code/
Author: Ani Björkström — QvantX Sweden AB, Stockholm
Published: 2026-05-01
Updated: 2026-09-20
Video: https://www.youtube.com/watch?v=M_lpQeRwP10 (8 min)
Topic: Careers

License: free to quote and cite with attribution to https://qvantx.com

---
_CAREERS_

## Can a No-Code AI Agent Really Find 2,002 Jobs For You?

#### Key takeaways

- A website change detector built in Claude Cowork, without writing code, monitors 69 job-listing URLs and has collected 2,002 jobs in its history.

- One location filter cut those 2,002 jobs down to 384 Stockholm roles, replacing a daily routine of opening 69 pages by hand.

- The full setup needs only an Anthropic API key, a Gmail app password created in Google Cloud, and Docker to run the app on your own computer.

Checking dozens of job boards daily is a chore that scales badly. Ani Björkström, a tech consultant based in Stockholm, replaced that routine by asking Claude Cowork to build a website change detector that scrapes 69 job-listing URLs, stores every posting in a searchable history, and emails a digest of new opportunities on a schedule.

### How do you build a job-finding agent without writing code?

You describe the app in plain language to Claude Cowork and answer the clarifying questions it asks before building. Claude responded to the prompt "create a website change detector" with key questions: how listings should be extracted, how email notifications should be delivered, how URLs get added, even which text editor is preferred.

Each question comes with a recommended option, so the whole specification is answered by multiple choice. While it builds, Claude shows its task list under a progress panel and writes real files to a folder on your computer, including a requirements file, environment example, gitignore, and Dockerfile.

### What choices does Claude ask you to make before it builds?

Claude asks about extraction, delivery, hosting, and error handling before writing any code. These were the answers behind the finished job agent.

| Setup question | Option selected |
|---|---|
| How should listings be extracted? | Listing URL plus title, the recommended option |
| How should email notifications be delivered? | One digest per schedule |
| Which LLM API provider for extraction? | Anthropic Claude API, on an existing subscription |
| Where should the app run? | A Docker container on the user's own computer |
| Which extra behaviors are enabled? | Retry plus error alerting when scraping fails |

### What do you need to configure before the agent runs?

Three things: an Anthropic API key in the .env file, a Gmail address as the SMTP username, and an app password created in Google Cloud Platform as the SMTP password. If Docker is not installed, that is the fourth and final prerequisite before running the single Docker command that starts the container.

Once the container runs, the app is served from your own machine at a local HTTP URL. After several iterations with Claude, the result was a working website with a URL tab, a run button, and a history view.

### How much time does the agent actually save?

Instead of visiting 69 URLs and reading every listing one by one, one click fetches all jobs and an email digest delivers the new ones. The history view held 2,002 jobs after three runs, and filtering by location reduced that to 384 Stockholm positions, each opening the original posting with its apply button one click away.

#### FAQ

**Do I need programming experience to build this agent?**

No. Claude Cowork asks multiple-choice questions, then generates every file itself, including the Dockerfile. Your work is answering questions and pasting credentials into the .env file.

**Where does the application actually run?**

In a Docker container on your own computer, which acts as the server. You access the app through a local HTTP URL once the container is started with a single Docker command.

**How do new job listings reach me?**

By one digest email per schedule, so you are not pinged for every posting. If scraping a site fails, the app retries and sends an error alert, because retry plus error alerting was enabled during setup.


## Full video transcript

[0:10] Hi, if you are trying to find a job, in this tutorial I will show the exact workflow I use to land on the best opportunities. If we haven't met, my name is Annie. I'm a tech consultant based in Stockholm. Let's get started. I use Cohere work and I ask Cohere work to create a website change detector for me. Here Cloud answers, I'd love to help to build this project. It's a solid idea, but before I start, let me ask some key question to make sure I build the right thing. And those are the questions. You can see the list of all the questions that Cloud want me to answer. The first one is, different job project sites have very different HTML structure. How should the app extract listing? Then it ask, how should the app send you email notification?

[0:58] And then the next question is about do you want for adding URLs? And then the last one, which text editor do you prefer? Now, for the first question, I will choose the recommended option by listing URL plus title. Next question, it asks, how should email notification be delivered? And there I will be choosing on digest per schedule because this is the recommended option as well. Then it asks, many modern job sites load content via JavaScript. How should the app handle those? And here I will choose the recommended one. And then the next question is, which fields should the app extract from each listing? Pick all that apply. And I will choose all of those because I want to extract as much information as possible from the websites.

[1:46] Then it says, great choices, a few more important questions. So it will ask even more questions. And here again you can see the list of the questions that it will ask and we will be given some choices and then you can choose one of those. So, the first question is, I extraction needs an LLM API, which provider you prefer? I will go with Anthropic Cloud API because I already have a subscription and I paid for that and I think it is reasonably priced and works fine. Next question is, who will use this app? I will go with just me, single user. And also, this is the recommended. Next one is for schedule trials. The app needs to be running somewhere. How? And here I will choose the first recommended option that is Docker container.

[2:31] And then the next one is a few extra behaviors enable any you want. And I want to enable in here retry plus error alerting, which means that if there is some type of errors while scraping jobs from websites, I want it to retry and then to alert that there was an error. I will choose this one in here and then I will click next. So, here you can see all the questions with our answers. Now, it is thinking and the great thing about Cloud is that we can see all the fault process. So, it says, "Perfect, I have everything I need. Let me verify the workspace and start building." On the left side under progress, you can see that it is creating different tasks that it needs to work on for the progress.

[3:19] And then underneath that, you can even see website change detector. And there, right now we have just instruction Cloud MD file, but more files will be added in there as well. So, again, under the progress are all the tasks that Cloud needs to work on. And then under website change detector folder, we can see all the files that will be created by Cloud Cowork. And website change detector actually is a catalog in our computer and we can even open it and see all the actual documents in there as well. You can see that we have requirements file, environment example, gitignore, instructions, docker file, and those files are being added while Cloud is working on the project.

[4:05] And if I go to the folder, and here you can see that we have the actual folder, and you can see files inside the actual folder. So, it says perfect. I have everything I need. Let me verify the workspace and start building. And then underneath that, you can see the full process, all the tools. It says, "Now, let me build the app. I will start the project." Now, Cloud created the application, but there are still some things that we need to configure. First one is to provide Anthropic API key, and we need to enter that information in env file. Next, we will provide our Gmail address in front of SMTP username, and next we will create an application in GCP, and we'll provide password in here in front of SMTP password.

[5:00] Next, if you don't have Docker installed in your computer, then you will need to install it. And after that, you can run the Docker command that is under step three in here that starts with Docker. This will initiate and run your Docker container, which means that your application will after that be available via your computer that is acting as your server. So, you can access it if you go to the HTTP URL that is in step after several iteration. We have a working website, and this is how it looks. You can see that I have run it three times, and I want to show you how it works. If I go to URL tab, you can see that I have 69 URLs, and those are all the URLs that I daily open to find all the job listing.

[5:49] For example, if I click on one of them, you can see that in here we have those three listings. So, what I normally will do if I don't use this application is going to one and each of those URLs and reading all the available job listings and then after that defining which ones are interesting for me and applying to those. However, this process takes a lot of time. In order to make this process easier, I created these websites that instead of me is checking URLs and getting all the job listings and saving in the history. So, if I go to history, you can see that there are 2002 jobs in here and if I find something that is interesting for me, I will just click on that and the job will open and then I will apply for that job.

[6:40] For example, in here we have system engineer role that is in Stockholm, so I will click on this and the job will open. Then I will click on this button to apply for this job. Here are many different opportunities and not all the jobs are in Stockholm. So, in order to filter the jobs that are in Stockholm, I will use the location and in here I will just choose Stockholm and will click on apply. So, all the jobs that are in Stockholm will pop up in here and instead of 2002 jobs, here we have 384 jobs. The reason this workflow is creating a lot of value for me is that instead of visiting all of those 69 URLs and getting the jobs one by one, I can click the button and the system will get all the jobs and send me an email with all the job opportunities.

[7:31] If you want to use the same web application and want to get the prompt I used to create it, comment prompt down below and I will send it to you for free. If you like this video, please give me thumbs up and see you in the next one. Bye.