9.2 KiB
GSOC Proposal for DashBot
About Me
Full Name: Mrudul Killedar
Email: mrudulkilledar111@gmail.com
Phone: +91 7489685683
Discord Handle: Mrudul (username: mk4104)
GitHub Profile: https://github.com/Mrudul111
LinkedIn: www.linkedin.com/in/mrudul-killedar-5b0121245
Time Zone: GMT + 5:30
Resume: https://drive.google.com/file/d/1ICvI5h8FP5cTtMIvQcu918Adc5JP1DQb/view?usp=share_link
University Information
University Name: Vellore Institute of Technology
Program: B.Tech in Computer Science Engineering
Year: 2025
Expected Graduation Date: July 2025
Motivation & Past Experience
1. Have you worked on or contributed to a FOSS project before?
Yes, I have contributed to FOSS project ie APIDash:
- Solves Beautify JSON and Highlight JSON - https://github.com/foss42/apidash/pull/595
- Share button functionality - https://github.com/foss42/apidash/pull/571#event-16324056931
- Homebrew Installation Guide - https://github.com/foss42/apidash/pull/566#event-16282262849
- Multiple-Model in DashBot - https://github.com/foss42/apidash/pull/704
2. What is your one project/achievement that you are most proud of? Why?
I am most proud of 2 projects particularly one is Parking24 which is an app that was made to solve parking problems of UAE market and i was assigned this project in my internship at Wisho (Now called Hyve AI Labs).
I am proud of the project because it is adding real world value and solving a very real problem. Also i have been working on a Machine learning model that is to predict cybersickness using SNN and currently we are achieving accuracy of 87.234% which is better than existing model that has accuracy of 76.11% and we achieved this because of our better approach for data cleaning.
3. What kind of problems or challenges motivate you the most to solve them?
I am most motivated by solving complex problems that challenge me to think in new ways and push my boundaries. I enjoy tackling problems that I have never encountered before, as they provide an opportunity to learn, explore innovative solutions, and develop a deeper understanding of different technologies. The thrill of breaking down a difficult problem, analyzing it from different angles, and coming up with an effective solution is what drives me the most.
4. Will you be working on GSoC full-time?
Yes, I will be working full-time on my GSoC project.
5. Do you mind regularly syncing up with the project mentors?
Not at all! I am happy to have regular sync-ups with my mentors to ensure smooth progress and alignment with project goals.
6. What interests you the most about API Dash?
API Dash is an innovative tool that simplifies API testing and monitoring. I personally felt other apps are very bloated and performace is really clunky on the other hand API Dash is really smooth, looks aesthetically pleasing and use of AI just seperates them from rest of the API Testing Platform. I have personally shifted to API Dash for my backend testing. I would love to contribute to API Dash because i feel this is a project that is adding such a great value to developer community.
7. Can you mention some areas where the project can be improved?
Some areas where API Dash can be improved include:
- UI for DashBot: The UI currently is very basic and lacks a professional approach.
- Responses: The response that is generated by clicking the buttons is on-point but the bot is not conversational enough.
Project Proposal Information
Proposal Title: DashBot
Abstract
DashBot is an AI-powered assistant designed to supercharge developer productivity within API Dash by automating repetitive tasks, improving API debugging, and providing intelligent recommendations. By leveraging advanced large language models (LLMs), DashBot enables developers to interact with APIs using natural language, making API testing, debugging, documentation, and integration significantly more efficient and intuitive.
Detailed Description
DashBot is designed to be an AI-powered assistant for API Dash that helps developers automate tedious tasks, follow best practices, and obtain contextual suggestions via natural-language input. First we need to finish tasks that are yet to be finished in the first prototype which include:
- Generate plots & visualizations for API responses along with ability to customize
- Generate API integration frontend code for frontend frameworks like React, Flutter, etc.
This project extends its capabilities by adding the following advanced features:
1. AI-Powered Code Error Detection & Auto-Fix
- Detect syntax and logical errors in API requests and integration code.
- Provide human-readable explanations and suggest one-click fixes.
- Ensure best practices in authentication, rate limiting, and error handling.
2. Multi-Model Support & Fine-Tuning
- Enable users to switch between different LLMs (GPT, Llama, Claude, Gemini, etc.).
- Provide on-device LLM support for private inference.
- Allow user-defined prompt fine-tuning for personalized suggestions.
3. Enhanced UI/UX for API Dash
Improve the overall user experience of API Dash by making the interface more intuitive, visually appealing, and developer-friendly.
- Modern UI Elements: Redesigned buttons, input fields, and layouts for a clean and professional look.
- Dark & Light Mode Support: Seamless theme switcher for better accessibility.
- Improved API Request/Response Visualization: Better syntax highlighting, collapsible sections, and JSON tree views for responses.
- Enhanced Error Debugging UI: Clear, structured error messages with AI-powered suggestions for fixes.
- Keyboard Shortcuts & Command Palette: Faster workflows with keyboard commands.
4. API Documentation for Tech Stack Integration
- Provide Step-by-Step Guides for React, Flutter, Vue, Express.js, and more and change the response according to coding practice used by the user.
- Allow Markdown and PDF export for easy sharing.
Packages Used
This project will utilize the following packages to implement the proposed features:
anthropic_sdk_dart- Claude integrationgoogleai_dart- Google AI model supportopenai_dart- OpenAI API accessollama_dart- OllamaAI inferencefl_chart- API response visualization
Figma Link
https://www.figma.com/design/vzpQ7xzwwmx2G92VVyF4aw/GSOC-Proposal?node-id=0-1&t=7D6Njm8Rr2x6VCkx-1
Weekly Timeline
Week 1: Initial Research & Understanding API Dash
Tasks:
- Study API Dash’s existing codebase.
- Research AI assistant implementations.
- Set up the development environment.
Week 2: Finalizing Tech Stack & Initial UI Prototyping
Tasks:
- Decide on AI models & packages to be integrated in DashBot.
- Finalize the UI wireframe.
- Create a separate development branch for DashBot features.
Week 3-4: UI/UX Enhancements
Tasks:
- Implement DashBot Panel UI.
- Add theme support (light & dark mode).
- Improve API response display with syntax highlighting.
- Implement API request history tracking and auto-suggestions.
Week 5: Frontend Code Generation
Tasks:
- Generate API integration code for React, Flutter, Vue.
- Allow one-click copy of generated code.
Week 6: API Response Visualization
Tasks:
- Implement data visualization for API responses.
- Add customization options for plots. Deliverable: Interactive API response visualizations.
Week 7: AI-Powered Code Error Detection
Tasks:
- Implement AI-powered debugging for API requests.
- Ensure API best practices compliance.
- Optimize AI model for faster debugging.
Week 8: Multi-Model Support & Fine-Tuning
Tasks:
- Implement model switching UI.
- Connect multiple AI models via API.
- Integrate on-device AI support.
- Enable custom prompt fine-tuning.
Week 9: Documentation for Integration of API to specific tech-stack
Tasks:
- Generate API integration documentation for multiple languages (Python, JavaScript, Java, Flutter, etc.).
- Create UI for selecting the tech stack and displaying relevant API documentation.
Week 10-11: Testing, Debugging & Optimizations
Tasks:
- Conduct unit & integration testing.
- Fix bugs & optimize performance.
- Gather feedback from mentors & community.
Week 12: Documentation, Final Touches & Submission
Tasks:
- Write detailed documentation & API reference.
- Create a demo video & presentation.
- Prepare & submit the final report to GSoC.