📄 Ticketeer Llama
Overview
Ticketeer is a customized Large Language Model based on Llama 3.1:8B, designed specifically to generate high-quality, structured issue from plain user input.
It automatically classifies incoming requests into one of five issue types:
- Bug (error, malfunction, defect)
- Feature Request (new capability or improvement)
- General Task (maintenance, update, administrative work)
- Epic (large initiatives requiring breakdown into smaller stories)
- Spike (time-boxed research, investigation, or prototyping tasks)
Key Features
- 🔥 Auto-Classification: No need to specify the ticket type manually — the model intelligently determines it based on input context.
- 🛠️ Structured Ticket Generation: Each ticket type has a tailored structure, including fields like Summary, Description, Priority, Components, Labels, and specific sections such as Steps to Reproduce, Acceptance Criteria, or Research Approach.
- 🧠 Reasoning Phase: Before generating the final ticket, the model provides a thought process wrapped in
<think></think> tags, showing classification reasoning, urgency assessment, recommended priority, and component analysis.
- ✨ Markdown-Formatted Output: Field labels are bolded for easy copy-paste into JIRA or documentation systems.
- 🔎 Missing Information Handling: If necessary details are missing, the model clearly flags them as “TBD” to ensure ticket completeness.
How It Works
User provides a simple description, such as:
“The user profile page crashes when clicking ‘Save changes’ with invalid data.”
The model:
- Internally classifies it as a Bug.
- Thinks about urgency, priority, and impact.
- Outputs a complete, well-structured issue, ready to use.
Tickets include sections appropriate to the type, e.g., Steps to Reproduce for bugs, Success Criteria for epics, and Research Approach for spikes.
Intended Use Cases
- Engineering teams drafting issues faster and more consistently
- Agile project management workflows
- Technical Product Owners and Scrum Masters standardizing ticket quality
- Automating intake from vague customer requests into actionable stories
Example Ticket Types and Structures
| Type |
Key Sections Included |
| Bug |
Observed Behavior, Expected Behavior, Steps to Reproduce |
| Feature Request |
Feature Description, Value, Acceptance Criteria |
| General Task |
Task Details, Reason/Context |
| Epic |
Epic Goal, Business Value, Success Criteria, Child Tickets Needed |
| Spike |
Problem/Question, Research Approach, Expected Outcome, Timebox |
Installation / Usage
After loading the model:
ollama run omaciel/ticketeer-llama3.1
Provide a description of a problem, feature, task, or investigation. The model will automatically produce a complete issue suggestion.
🚀 Ready to transform vague ideas into actionable, professional issues — in seconds.