At Team '26 Europe, Atlassian introduced AMP, a framework giving agents an assigned identity, scoped authority, shared Teamwork Graph context and reviewable output, and made a rebuilt Atlassian MCP generally available with 220+ tools, OAuth 2.1/PKCE and tiered read/write/destructive scopes.

Key Takeaways

  • ✓Atlassian MCP handles 15M+ calls/day; rebuilt from dozens of tools to 220+, with up to 25% token savings on the same Jira/Confluence work in internal Claude benchmarks
  • ✓Teamwork Graph (250B+ connections) now indexes Bitbucket/GitHub source down to functions, symbols and classes via Rovo Code Search, no clone needed
  • ✓Internal benchmark: agents grounded in Teamwork Graph were 44% more accurate with 48% fewer tokens
  • ✓Humans and agents already collaborate 10M+ times a month on Atlassian; Jira agent sessions link local and cloud agent runs back to work items
  • ✓MCP secured with OAuth 2.1 + PKCE by default, read/write/destructive scopes separated; EU-local AI inference added, non-human identity inventory coming soon
Atlassian launches Agentic Multiplayer Protocol (AMP) and a rebuilt Atlassian MCP with 220+ tools, 15M+ daily calls and up to 25% fewer tokens
🖼️Official Media
Click to view high-res
⚡

Key Decision Metrics at a Glance

TL;DR Verdict
🏆 Benchmark RankRank
Atlassian MCP handles 15M+ calls/day; rebuilt from dozens of tools to 220+, with up to 25% token savings on the same Jira/Confluence work in internal Claude benchmarks
Full Leaderboard→
💰 Cost & PlansCompare
API & Subscription
Compare Plans↗
🛠️ Ecosystem & IDERules
Rovo · Claude Code · Cursor
Setup IDE Rules↗
🧭

Heavy Claude Code use: compare subscription limits and API bills

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

⚔️ Compare Model→💰 Compare plans & limits↗
🔗

Project Links & Resources

Direct Access
Direct access to official project resources and documentation
📝
The Atlassian platform is for multiplayer human-agent work (AMP announcement)
www.atlassian.com/blog/company-news/team26-europe-ai-platform
↗
📝
Atlassian MCP brings more of your work within reach of any AI agent
www.atlassian.com/blog/company-news/team26-europe-atlassian-mcp
↗
📝
Introducing Code Context: AI context engine for your codebase
www.atlassian.com/blog/development/code-context
↗
📖
AMP overview - Atlassian Design
atlassian.design/amp
↗
🔬

In-Depth Technical Analysis

Atlassian used Team '26 Europe to announce Agentic Multiplayer Protocol (AMP), its framework for humans and agents working under the same rules: each agent gets an admin-assigned identity, scoped authority, shared tasks and context from the Teamwork Graph (250B+ connections), and must return reviewable results. The Teamwork Graph now indexes Bitbucket and GitHub source down to functions, symbols and classes (Rovo Code Search) and adds data lakes and warehouses as structured context. The rebuilt Atlassian MCP is generally available with 220+ tools across Jira, Confluence, Bitbucket, Loom and Goals, toolset-level scoping, separate read/write/destructive tiers and OAuth 2.1 + PKCE by default. Atlassian reports 15M+ MCP calls per day and up to 25% token savings versus the previous version in internal Claude benchmarks; its earlier Code Context post claimed 44% higher accuracy and 48% fewer tokens for graph-grounded coding agents. All figures are vendor-reported. Jira also gains agent sessions that link local and cloud agent runs back to work items, plus EU-local inference; non-human identity inventory is still listed as coming soon.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

⚔️ Launch VS Compare→💳 View Plans↗