This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle
🔨 What I Built
Have you ever agreed to build a website for a fixed price, only for the client to keep adding “just one more feature”?
That is scope creep. For freelancers, it can mean unpaid hours, missed deadlines, difficult conversations, and projects that become far bigger than originally agreed.
I built ScopeSentry, an AI-powered scope analysis tool that helps freelancers understand what changed between their original client agreement and a new request—before they say yes.
ScopeSentry compares two pieces of information:
The original agreement: What the freelancer and client initially agreed upon.
The new request: What the client wants to add or change.
It analyzes the differences and produces a structured breakdown designed to help freelancers make informed decisions.
Key features include:
Scope change detection: Identifies requested features that may fall outside the original agreement.
Evidence-based analysis: Uses quotes from the original agreement to support its findings.
Effort and cost estimation: Estimates additional work in hours and calculates a potential cost impact using the freelancer's hourly rate.
Uncertainty detection: Highlights missing details and areas that need clarification instead of treating every assumption as fact.
Suggested client reply: Generates an editable response to help freelancers discuss additional requirements professionally.
The goal isn't to automatically reject client requests. It's to help freelancers understand the implications before agreeing to them.
🎬 Demo
Watch ScopeSentry in action: https://youtu.be/G4vH6z9WxTU
The demo walks through a website-development agreement where a client subsequently requests a blog, an admin dashboard, user authentication, and a customer profile page without changing the original budget or deadline.
ScopeSentry analyzes the new request and presents the potential scope changes, supporting evidence, estimated effort, cost impact, and a suggested client response.
**
💻 Code**
GitHub repository: https://github.com/DevSrivastava14/SCOPE_SENTRY.git
The project uses a React + Vite frontend and a Python FastAPI backend, with AI-powered analysis through the Groq API using an open-weight language model.
⚙️ How I Built It
I built ScopeSentry with a focus on making AI analysis useful, structured, and reviewable rather than simply returning a block of generated text.
Frontend — React + Vite
The frontend provides a form for entering the original agreement, the client's new request, and the freelancer's hourly rate. It presents the analysis in a readable format and allows the suggested client reply to be edited and copied.
Backend — Python + FastAPI
The backend validates incoming requests, communicates with the model provider, parses the model's output into a structured response, and checks that evidence quotes are grounded in the original agreement. It also handles malformed responses and provider errors.
AI — Open-weight language model
ScopeSentry uses openai/gpt-oss-120b through the Groq API for agreement comparison and analysis. The model helps interpret the two texts, identify potential additions, and produce structured findings.
I also added automated backend regression tests and checked frontend linting, production builds, and important UI behaviors. The implementation aims to make failures easier to detect and the results easier for users to review.
🌍 Why Open Innovation Matters
Scope creep is a common problem, but freelancers do not always have access to expensive project-management tools, contract specialists, or dedicated account managers.
Open-source AI creates an opportunity to build practical tools for people who might otherwise be underserved by enterprise software. Developers can inspect the implementation, learn from the approach, adapt it to their workflows, and contribute improvements.
ScopeSentry is a small example of that potential: combining an accessible web stack with an open-weight model to tackle a specific, everyday business problem.
There is also an important responsibility here. AI-generated estimates are not guarantees, and a model should not invent contractual terms or replace a freelancer's judgment. ScopeSentry is designed to support human decision-making, not make it on the user's behalf.
I hope this project can evolve through community contributions, better evaluation cases, improved evidence validation, and feedback from freelancers who experience scope creep firsthand.
**
🤖 My Agent Session
**
I used GitHub Copilot as a coding assistant during development. I worked through the project incrementally, using it to help implement the frontend and backend, troubleshoot integration issues, and strengthen automated testing.
The process also reinforced an important lesson: building an AI application is not just about connecting a model to a form. Input validation, evidence grounding, error handling, configuration, and testing all matter when people rely on the output.
What's Next?
Potential improvements include user-specific usage limits, stronger evaluation against real-world agreement examples, support for additional document formats, and optional integrations with freelance project-management workflows.
For now, ScopeSentry focuses on one clear goal:
Help freelancers recognize scope changes, understand their potential cost, and have better conversations with clients.
Team Submissions: Made the Project solo username : devsrivastava14
I'd love feedback from freelancers and open-source developers: What would make you trust a tool like this when a client asks for “just one more feature”?
Top comments (0)