This article details how to integrate a local Large Language Model (LLM) with Jira projects using LM Studio. The author demonstrates setting up a local Qwen/Qwen3.6-35B-A3B model on a Mac Studio to act as a validator for specific Jira Cloud projects, while other projects continue to use the default cloud provider. The setup involves exposing the local LM Studio server via Tailscale Funnel and configuring CogniRunner, a Jira app, to route requests to the local model. The author tested the system by attempting to submit a request with a sensitive phrase, which was correctly blocked by the local LLM, and also tested a fail-open scenario by providing an incorrect API token, which resulted in the request bypassing the check. AI
IMPACT Enables localized, project-specific AI validation within Jira, offering enhanced privacy and control over sensitive data processing.
RANK_REASON The article describes a specific technical integration and setup for using a local LLM with a project management tool, which falls under tooling.
- Apple M3 Ultra
- Atlassian
- CogniRunner
- Jira
- LM Studio
- Mac
- Mac Studio
- OpenRouter
- Qwen/Qwen3.6-35B-A3B
- Tailscale Inc.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →