Launching Another AI Startup with No-Code (and looking for beta users)
Back story
It all started several months ago, when I came up with an idea of creating an AI startup in just one weekend. SummaLetter started as a weekend project aimed at solving the problem of newsletter overload. Using Bubble I was able to build and launch SummaLetter in just one weekend. The tool automatically summarized newsletters, making it easier for users to digest information quickly. I described this story on IndieHackers and it quickly went viral leading to 400+ users.
A Successful "Exit"
As part of the Product Launchers Club, I sold SummaLetter for $1,500. The Club is a unique community dedicated to transforming trending micro-startup ideas into market-ready products using no-code tools. The sale of SummaLetter validated the potential of no-code solutions and micro-SaaS projects, proving that small exits can be meaningful for me.
The New Idea Comes to Life
Every month, the Product Launchers Club members picks a trending product idea, which is then built live during a webinar. CraftYourJob was one such idea that got the green light. I embarked on turning this concept into a real, functioning product.
Preparation Phase
Not to start from scratch I chose a free landing page from Figma templates, which offered a variety of options that could be customized to fit our brand's aesthetic. This was crucial because it allowed me to rapidly set up a presentable and functional page to welcome our users.
With the landing page up, I turned to Bubble, which is always my first choice to build apps. Bubble's interface, split into sections for design, workflows, data, files, and plugins, is intuitive yet powerful, perfect for both beginners and experienced builders. My first task on Bubble was to create simple layouts that would form the foundation of our application. These initial setups were crucial for testing ideas and making quick iterations based on feedback.
The Webinar: Implementing Core Functionalities
The heart of CraftYourJob lies in its ability to let users compare their resumes with job descriptions. This process was divided into two main functionalities I've built during the 1-hour live webinar prepared for Club members and some guests.
PDF Resume Upload
To enable users to upload their resumes, I integrated a feature that not only accepts PDF files but also converts these files into text. This was achieved using a plugin available on Bubble called "PDF Text Extraction." Here’s how I implemented it:
- Extraction Workflow: Once a resume is uploaded, the application triggers a workflow in Bubble that utilizes the "PDF Text Extraction" plugin. This plugin works by reading the PDF file and extracting all the text content from it. I configured the plugin to handle multiple pages within a single PDF, ensuring that no content is missed, regardless of the resume's length.
- Data Handling: After the extraction, the text data is temporarily stored within the application’s state. This data handling is crucial as it allows for immediate processing in the subsequent steps, particularly when comparing the resume text to job descriptions.
Job Description URL Input
Another core functionality is the ability for users to input the URL of a job description they are targeting. This process involves capturing a URL and retrieving the text from the job description web page:
- Web Scraping: To extract text from the given URL, I used Bubble's capability to integrate external APIs. In this case, I employed a scraping tool configured within Bubble to fetch and parse the HTML content of the web page associated with the URL. This tool is adept at navigating through HTML structure to locate and extract the job description, ignoring irrelevant content like menus, footers, and ads.
- Text Processing: The extracted text is then cleaned and formatted to remove any extraneous HTML tags or scripts that might have been fetched during the scraping process. This ensures that the final text used for comparison is clean and relevant to the job application.
Leveraging AI with OpenAI Plugin
With the core functionalities set up, the next significant step was to integrate AI to provide deep insights and comparisons between the resume texts and job descriptions. This integration was facilitated through OpenAI plugin, which connects our no-code application to one of the most advanced AI models available today in just a few minutes.
AI-Powered Analysis
- Text Summarization and Analysis: Using the capabilities of OpenAI’s GPT model, the application requests AI to perform text summarization on both the resume and the job description independently. This AI-driven summarization condenses the content into key themes and qualifications, making it easier to compare them directly.
- Comparative Analysis: Beyond summarization, I programmed the AI to compare these summaries to evaluate how well the user’s qualifications matched the job requirements. This comparison is not just a direct text match; it involves understanding the context and relevance of the skills and experiences mentioned in the resume relative to the job description.
- Output Generation: Based on the AI’s analysis, the application generates a match score and detailed feedback on areas of improvement. This feedback might include suggestions on skills to develop, certifications to obtain, or experiences to gain, all tailored to increase the user's job compatibility.
Real-Time Feedback to Users
Once the AI completes its analysis, the results are immediately ready to display to the user within the application. This includes both the match score and the detailed suggestions for improvement, all rendered in an easy-to-understand format that encourages user engagement and action.
Launch and Beyond
So, I started building on Friday, went online on Sunday. As CraftYourJob has just launched, I'm actively looking for beta testers to provide valuable feedback and help refine the tool.
I'd love to discuss your ideas and insights both on the idea, and no-code movement in general. And let me know if you'd like to have more projects described like this!
