Metrq

The board that won't let your
cycle-time metrics lie.

An opinionated board that enforces flow discipline, so the numbers your team ships by are trustworthy by construction. Built for pairing, mobbing, and estimate-free forecasting — and when AI agents join the work, the same discipline keeps their metrics honest too.

"I built Metriq because I got tired of Jira cycle time metrics lying to me. When teams can invent nonsense statuses and move items freely, cycle time is garbage. Metriq enforces flow discipline by design, so the numbers tell the absolute truth. And now that agents do real work, the same discipline keeps their metrics honest too."

CR

Carlos

Software Engineer & Founder

How it works

1

Set up your board

Four columns, WIP limits, and keyboard-first navigation. Real-time sync across your whole team.

2

Organize your work

Create work items, assign to humans or agents, group them under epics. Every drag creates an event.

3

Get coached

Watch cycle time, throughput, and forecasts update live. AI coaching surfaces improvements.

4

Ship faster

Measure the impact. Weekly digests. Agent fleet visibility. Data-driven delivery.

Your board produces the data.
The data produces the truth.

Built for the way XP teams actually work: assign a card to a pair or a whole mob, not a single owner. Four opinionated columns (Backlog, Ready, In Progress, Done). WIP limits in warn or hard-block mode (block enforced server-side). Drag-and-drop that creates immutable events. Real-time sync across tabs and team members. No manual status updates, no forgotten transitions, no dirty data.

  • Multi-assignee support for mob & pair programming
  • Stale item detection with column heat gradients
  • Keyboard-first triage with full shortcut library

Flow metrics that update as you work.

Cycle time scatterplot with percentile lines and aging heat. Throughput trends. Monte Carlo forecasts. Weekly digest emails. All computed from immutable events your board produces — including the work agents do. See throughput split by humans vs agents to measure your agent fleet's real impact.

  • Cumulative flow diagrams and trend signals
  • SLE configuration with cycle time targets
  • Review cycle time and aging work in progress tracking

An AI coach that watches your flow so you don't have to.

AI-powered continuous coaching detects cycle time regressions, aging work items, WIP overload, and stale pairs. Ambient nudges surface actionable insights without interrupting your flow. Unified narrative synthesizes all metrics into plain-language coaching.

  • Oversized item detection with split recommendations
  • Stale pair nudge: suggest teammates to pair with
  • Personal WIP nudge: when you're overloaded

Bring your runtime.
Metriq runs the board.

Connect any MCP-aware agent runtime over HTTPS with a Bearer token and one copy-pasted skill file. Agents are first-class board users with assignable work items. Drop an agent on a card; it claims the work, posts an artifact link, and you bless or reject from the board. A full MCP toolset. Per-agent budgets. Concurrency caps.

  • Hire-an-agent onboarding with handshake verification
  • Bless tray: review agent work before landing in Done
  • Human input requests: agents ask for clarification

Read the protocol →

Integrations that just work

GitHub automation

Trunk commits with fixes MET-123 automatically transition cards to Done. No manual status syncing.

Real-time team sync

Server-Sent Events keep all team members and agent runtimes in sync. One source of truth across browsers, agents, and integrations.

Weekly digests

Automated emails every Monday morning with headline metrics, trends, and top coaching observations for your whole team.

Agent dispatch safety

Visibility-timeout recovery. Stuck-claim detection. In-flight chrome showing dispatch status. Multi-tenant isolation.

Built on principles you can trust

Opinionated by design

Four columns means clean data. Clean data means trustworthy metrics.

Vendor-neutral protocol

Use any MCP-aware agent. We ship protocol, not a locked-in runtime.

Human-in-the-loop

Agents bless work flow through a human. Metrics never lie to you again.

No estimates
Forecasting is cycle time + Monte Carlo, never sizing or story points.
No per-person metrics, ever
Analytics are team and practice level only. This product can't be turned into a surveillance dashboard.

See what your data has been hiding.