Why We Built xpntl

Issue trackers were designed for humans. But your team now includes coding agents that ship code at 3 AM. We built xpntl because the coordination layer between humans and AI doesn't exist yet.

The Problem

Every engineering team we talked to had the same story: they adopted Cursor, Claude Code, or Codex, their throughput jumped, and their existing tracker became the bottleneck. Agents can't navigate Jira. They can't parse a Linear board without a custom integration that breaks every quarter.

The result? Engineers spend more time translating between the agent and the tracker than the agent saves them. That's not exponential — it's a net negative.

What We Believe

A modern issue tracker needs to treat AI agents as first-class team members, not API consumers. That means:

• Native protocol support. MCP isn't an afterthought — it's the primary interface. Agents get the same 28 tools a human gets.
• Agent-aware workflows. When an agent picks up an issue, moves it to In Progress, and posts a comment — the board should reflect that in real time, just like a human.
• Transparent coordination. Every action is attributed. You always know who did what — human or machine.

The Name

xpntl is shorthand for exponential. Your agents already went exponential. Now bring your team. That's the gap we're closing.

What's Next

We're shipping fast. The MCP server, CLI, and web app are live. Self-hosting docs are coming. If you're running coding agents on a real team and your tracker can't keep up, grab an invite and try it.