OVERALL RATINGS
INSTALLS
2
SUPPORT
- Partner Supported
Key highlights of the appAI workflow automation for Jira — validators, conditions and post-functions, with your own AI key or zero-key Forge LLM
Validate meaning, not just structure
CogniRunner reads the meaning of your fields, attachments and issues, then blocks a transition when content fails your plain-English rule — and searches your project with JQL to catch duplicates.
Your AI, your key — local or cloud
Bring your own key for OpenAI, Anthropic, Azure, OpenRouter or Bedrock — or run models locally with LM Studio, or use the zero-key Atlassian Forge LLM. Switching providers never loses your settings.
Automate after the transition
After a transition, AI post-functions act: a semantic rule reads a field and writes another; a static rule runs AI-generated JavaScript at zero AI cost. Chain up to 50 steps — add comments, sub-tasks, or documents.
Supporting media
More details
CogniRunner brings real AI understanding to Jira workflows. Native validators only check structure; CogniRunner reads the meaning of
fields, attachments and issues — then validates, decides and acts, all in plain English.
- Validators & Conditions: block or hide a transition, with the AI's reasoning shown.
- Post-functions: semantic (read a field, write another) and static (AI-written code that runs at zero AI cost); plus comments, sub-tasks
and generate-and-attach DOCX/PDF/PPTX. - Reads attachments (images, PDF, Office) and runs agentic JQL to catch duplicates.
- Bring your own AI: Anthropic, OpenAI, Azure, OpenRouter, Bedrock — or local LM Studio, or the zero-key Atlassian Forge LLM.
- MCP tools: context7 (live docs), web-search, doc-processor; CogniRunner brokers every call.
- Skills, Memories and a Documentation library ground every generation.
- Admin hub: roles & scopes, execution logs, Add Rule wizard.
Open source on Atlassian Forge (Apache-2.0).
Resources
App documentation
Comprehensive set of documentation from the partner on how this app works
Privacy and Security
Privacy policy
Atlassian's privacy policy is not applicable to the use of this app. Please refer to the privacy policy provided by this app's partner.
Partner privacy policySecurity program
This app is not part of the Marketplace Bug Bounty program.
Integration permissions
CogniRunner integrates with your Atlassian app
Version information
Version 3.0.0•for Jira Cloud
- Release date
- Jun 18th 2026
- Summary
- Post-functions, Forge LLM, LM Studio, BYOK Multi provider
- Details
A major update — from AI validation to a full AI automation platform.
- AI post-functions: semantic (read a field, decide, write another) and static (AI-written code that runs at zero AI cost). Plus
comments, sub-tasks, and generate-and-attach DOCX/PDF/PPTX. - Bring your own AI: Anthropic, OpenAI, Azure, OpenRouter, Bedrock — or local LM Studio, or the zero-key Atlassian Forge LLM. No embedded
key. - MCP tools: context7 (live docs), web-search, doc-processor — CogniRunner brokers every call.
- Agentic validation: autonomous JQL across the issue's project, plus attachment reading (images, PDF, Office).
- Skills, Memories and a Documentation library ground every generation.
- Admin hub: roles & scopes, execution logs with traces, Add Rule wizard.
- Hardened: sandboxed code, prompt-injection fencing, durable async jobs. Refreshed UI with dark mode.
- AI post-functions: semantic (read a field, decide, write another) and static (AI-written code that runs at zero AI cost). Plus
- Payment model
- Paid via Atlassian
- License type
- Apache License, Version 2.0 (ASL)