AI-assisted developer tool that extracts structured tickets from screenshots (Jira, Azure DevOps, GitHub, etc.), validates them against schemas, and generates a clean, normalized ticket ready for development workflows.
Ticket Forge removes the friction of manually reading tickets from screenshots, Slack messages, or documentation by transforming them into structured and validated data automatically.
Developers frequently receive tickets through:
- screenshots
- Slack messages
- documentation
- project management tools
These inputs often contain incomplete or inconsistent information.
Ticket Forge uses AI + schema validation to:
- Analyze ticket screenshots
- Extract structured fields
- Normalize values
- Detect missing critical information
- Present a review interface
- Generate a clean normalized ticket
flowchart TD
User[Developer Uploads Ticket Screenshot]
User --> UI
UI[Next.js Frontend Interface]
UI --> API
API[Next.js API Route]
API --> AI
AI[OpenAI Vision Model]
AI --> Extraction
Extraction[Ticket Field Extraction]
Extraction --> Validation
Validation[Zod Schema Validation]
Validation --> Normalization
Normalization[Ticket Normalization Layer]
Normalization --> MissingFields
MissingFields[Missing Field Detection]
MissingFields --> Review
Review[Human Review UI]
Review --> FinalTicket
FinalTicket[Normalized Universal Ticket Output]
The developer uploads a screenshot of a ticket from:
- Jira
- Azure DevOps
- GitHub Issues
- documentation or Slack
The screenshot is sent to an OpenAI Vision model which extracts:
- title
- description
- story points
- priority
- assignee
- status
- acceptance criteria
The AI returns a structured JSON response.
The response is validated using Zod schemas to ensure:
- required fields exist
- types are correct
- values match expected enums
Invalid or missing fields are flagged.
Raw AI outputs are normalized into consistent values.
Example:
| Raw Value | Normalized |
|---|---|
| P1, Urgent, High | High |
| Closed, Done, Resolved | Done |
Ticket Forge detects potential issues such as:
- story points not visible
- acceptance criteria embedded in description
- assignee unclear due to cropping
These insights are shown to the user.
Before confirmation, the user can:
- edit fields
- fix missing data
- confirm the ticket
This ensures AI + human reliability.
The result is a standardized ticket structure ready to be used by APIs or other tools.
type UniversalTicket = {
source: "jira" | "azure" | "github" | "unknown";
ticketId?: string;
title?: string;
description?: string;
priority?: string;
storyPoints?: number;
assignee?: string;
status?: string;
acceptanceCriteria?: string[];
};Analyze screenshots and extract:
- title
- description
- priority
- story points
- assignee
- status
- acceptance criteria
All data is validated using Zod schemas.
Checks include:
- required fields
- field types
- valid enumerations
- missing critical fields
AI outputs are standardized into consistent values.
The system automatically identifies incomplete ticket information.
Users can review and edit extracted tickets before confirming.
Frontend
- Next.js
- React
- TypeScript
- TailwindCSS
AI
- OpenAI Vision Models
- Structured JSON extraction
Validation
- Zod schema validation
Developer Tooling
- ESLint
- TypeScript strict mode
1️⃣ Upload ticket screenshot
2️⃣ AI extracts ticket fields
3️⃣ System validates extracted data
4️⃣ Missing fields are detected
5️⃣ User reviews ticket
6️⃣ Final normalized ticket is generated
{
"source": "jira",
"ticketId": "PHARM-29512",
"title": "Rename medication reminder creation flow",
"priority": "High",
"storyPoints": 5,
"status": "In Progress"
}Clone the repository
git clone https://github.com/AdrianaAC/ticket-forge
Install dependencies
npm install
Run development server
npm run dev
Open the application
Ticket Forge is currently under active development.
Planned improvements:
- multi-image ticket analysis
- acceptance criteria extraction
- Jira API integration
- GitHub issue generation
- AI confidence scoring
- batch ticket processing
Developers constantly deal with poorly structured ticket information.
Ticket Forge aims to:
- reduce manual interpretation
- improve ticket quality
- accelerate developer workflows
- demonstrate AI-enabled frontend architecture
Adriana Alves
Frontend Developer focused on AI-enabled developer tools and modern frontend architectures.
GitHub
https://github.com/AdrianaAC
LinkedIn
https://linkedin.com/in/adrianaalves098
⭐ If you find this project interesting, consider starring the repository.