Now in private beta

Your agentic PM.
Lives in Claude.

Roadmap visibility, sprint execution, QA, and infrastructure — all from the conversation you're already in. No dashboards. No context switching.

Get early access See how it works

Free forever · No credit card required

Claude — Patrick workspace
scenario 1 of 3
// what patrick does

Everything a PM does.
Inside the conversation.

Always know where things stand

Ask Patrick for sprint status, what shipped this week, or what's blocked. It reads your git history and backlog — git history, YAML backlog, sprint manifests — and tells you exactly where things are.

Patrick surfaces this automatically after every sprint ships. No manual updates.

Reads from git history + your codebase · no manual updates
U
Show me sprint status
P
Patrick sprint rollup ──────────────────── ● railway-deployment 91% HIGH ● insights-tab 57% HIGH ● stripe-billing 0% UNBLOCKED ──────────────────── 3 active · 2 unblocked

Know what to build next

Patrick reads your entire backlog, checks dependencies, estimates effort, and proposes the next sprint — ranked by priority and what's actually unblocked right now.

Before execution starts, Patrick assembles the full brief: existing capabilities, established patterns, relevant lessons. No re-inventing, no drift.

Checks what already exists before proposing anything new
U
What should we build next?
P
suggest_sprints("prioritize revenue") 1 · stripe-billing-integration 4-6h · high confidence · unblocked 2 · insights-tab-analytics 6-8h · high confidence · unblocked

QA that thinks

Three tiers: light (assertions, ~2 min), medium (full crawl + visual scoring, ~15 min), heavy (journey replay + simulation). Run any tier from Claude.

Patrick simulates real users through your product before your users do — finding broken buttons, UX friction, and visual regressions automatically.

Runs automatically after every sprint ships
U
Run QA on the dashboard
P
run_qa("medium") ✓ brief-preload PASS 1.2s ✓ groom-apply PASS 3.1s ✓ claim-toggle PASS 0.9s ✓ visual score 78/100 ────────────────────── 4 PASS · 0 FAIL · clean

Infrastructure without leaving Claude

Patrick's claws handle the tedious setup. Register a domain, deploy to Railway, wire up Stripe — all from the conversation. Each action is recorded so Patrick knows what's been provisioned.

Domain · Hosting · Payments · Auth · Email · Monitoring
U
Register buildwithpatrick.com
P
check_domain("buildwithpatrick.com") ✓ Available · $12.98/year register_domain("buildwithpatrick.com") ✓ Registered · expires 2027-06-08 DNS propagating — Patrick

Built for teams. Works solo.

Patrick coordinates claims, handoffs, and visibility across operators. Claim a workstream, see what's in flight, hand off without losing context.

Viewer seats are free. Bring in your PM, your co-founder, your advisor — they see what's being built without needing to touch the code.

Free viewer seats · Shareable sprint URLs · Shared context
U
What's Adrian working on?
P
get_sprint_status() Adrian · active since 10:30am claimed: stripe-billing/webhook status: in-flight · est. 2h You · suggested: insights-tab unblocked · 4-6h
// install in Claude Code
{ "mcpServers": { "patrick": { "command": "npx", "args": ["@patrick-ai/mcp"], "env": { "PATRICK_TOKEN": "your-token-here" } } } }

One config line · Works with Claude Code, Cursor, and any MCP-compatible client

// why it matters

Four problems.
One solution.

01

Visibility

"I have no idea what the AI built last week."

Patrick makes your entire build history readable — what shipped, what's in flight, what's next — surfaced in Claude on demand.

02

Memory

"Every session starts from scratch."

Patrick persists everything across sessions — what was built, why it was built, what to avoid. git knows history. Patrick knows history, present, and future.

03

Trust

"Claude rebuilt something we already have."

Before execution, Patrick checks what already exists in your codebase. No re-implementing. No pattern drift. QA runs automatically after every ship.

04

Coordination

"Who's building what? What's blocked?"

Patrick coordinates claims, handoffs, and visibility across operators and AI agents. Works solo. Scales to a startup team.

// pricing

Start free.
Pay as you grow.

Like DigitalOcean — remove the barrier to start, let usage drive the revenue. Free forever for small projects. Scales naturally as your product grows.

Free
$0 / forever
No trial. No expiry. Start building today.
  • Read-only MCP (roadmap, status, reports)
  • Up to 50 backlog entries
  • Light QA tier
  • 1 operator seat
  • Unlimited viewer seats
Pro
$19 / month
Full execution. Scales with your project.
  • Full MCP access — capture, execute, QA
  • Up to 500 backlog entries included
  • Medium + heavy QA + journey intelligence
  • Pattern + capability injection on every sprint
  • Daily and weekly reports
  • +$10/month per additional operator seat
Usage
Pay as you go
On top of any plan. You only pay for what you use.
  • Backlog entries above plan limit → per 1K/mo
  • Workstream runs above monthly limit → $1–3/run
  • Claw tasks (domain, Stripe, Railway) → per use
  • Heavy QA runs → per run
  • Affiliate revenue on every infrastructure setup

A solo vibe coder pays $19/month. A funded startup running Patrick seriously pays what it's worth. No sales call needed.

Start building
with memory.

Free forever to start. No credit card. Patrick is building itself with Patrick.

Free forever to start · No credit card required

✓ You're on the list. We'll be in touch.