Can we connect Jira with Internal AI tools

[Sudarsana Goddellagala]

We Have internal AI Tool which i am trying to connect with Atlassian for we have used MCP from Internal AI tool and i have added below information 

• I have created an Atlassian OAuth 2.0 (3LO) app
• I added Jira API scopes in the Permissions tab
• Then configured the callback URL in the Atlassian console
• In AI Tool MCP setup:
  • I have set Server URL  ( https://mcp.atlassian.com/v1/mcp/authv2)
  • I have  configured Authorization URL ( /auth/connect )
  • I have added Token URL ( https://auth.atlassian.com/oauth/token) 
  • I have  set Scopes ( read:jira-work,read:jira-user,offline_access )
  • I have  enabled PKCE
• triggered the flow using browser and “Connect & Discover Tools”  It Says Validation failed  for connect
• when I tried in Browser it asked me to select the Atlassian i have selected Sandbox and in next step 

  Authentication Failed

  Missing PKCE verifier in state.

Can some one help or share your experience if you connected with Cloud Jira with Internal AI Tool or MCP  

[Martin Runge]

Hi @Sudarsana Goddellagala

I would look closely at your Model Context Protocol (MCP) server configuration because the "Missing PKCE verifier in state error" specifically means your internal AI tool's authentication client dropped the temporary cryptographic key during the callback exchange.

• Are session cookies or local state memory enabled?
• You could try to switch your Atlassian app configuration to a standard Client Credentials grant flow instead: https://support.atlassian.com/user-management/docs/create-oauth-2-0-credential-for-service-accounts/