Built on Open Wallet Standard · MCP Native

Give every dollar a rule.
Give every role a boundary.

Policy-gated allowances for families teaching financial literacy and teams governing AI spend. One MCP server. Five roles. Zero key management.

Connect to Claude ▶ Watch demo
AllowanceAgent connected
Set up Maya with $15/week. 50% reading, 30% sports, 20% art.
Done. Maya's allowance is configured — $7.50 reading, $4.50 sports, $3.00 art per week. Savings: 20% auto-locked with 1.5× multiplier. Want to invite anyone else to the family?
Maya finished her reading goal — 90 out of 100.
Verified. Maya earned $6.75 in reading this week. Streak: 3 weeks 🔥. $1.35 auto-locked into savings. Ready to distribute?
12
MCP Tools
196
Tests Passing
5
RBAC Roles
0
Keys to Manage
Two Markets, One Engine
The same policy engine. Different vocabulary.

A parent says "50% reading, 30% sports, 20% art." A team lead says "60% AI tools, 25% compute, 15% training." AllowanceAgent doesn't know the difference.

📚

Achievement-Gated Allowance

Maya finishes her reading goal. The policy engine verifies, calculates her earnings, and distributes USDC — automatically. No manual approval.

🏦

Savings That Teach

20% auto-locks into a savings vault. After 30 days, it releases at 1.5×. Your child watches compound growth happen with real money, not a quiz.

👨‍👩‍👧

Whole Family, Five Roles

Invite grandma as family (gift-only), your co-parent (verify but not spend), or your daughter herself (view her own progress). Phone-speakable invite codes.

🏅

Gold-Backed Savings

Convert a portion of savings to PAXG (gold). Teach asset diversification: "USDC earns streaks, gold earns price appreciation." Real financial mechanics.

🤖

AI Spend Governance

Give each engineer $50/month — 60% AI subscriptions, 25% compute, 15% training. The policy engine enforces the split before the spend, not after.

📊

Automated Audit Trail

Every budget allocation, outcome report, and distribution logged with actor, timestamp, category, and amount. Finance pulls the audit log, not a spreadsheet.

🔑

Five Roles, Zero Overhead

Manager configures budgets. Team lead verifies outcomes. Employee reports usage. Advisor audits. No dashboards to build — team talks to Claude, ChatGPT, or Cursor.

🌐

Any MCP Client

Works with Claude, ChatGPT, VS Code, Cursor — any MCP-compatible client. Your team uses whatever AI tool they already have. AllowanceAgent doesn't care which one.

Access Control
Five roles. Cryptographically enforced.

Not software-level permissions. OWS policy-gated signing rules. Each role is a scoped API key with spend limits, wallet access, and tool restrictions baked into the signing layer.

👑
Manager
Full access. Configures policy, distributes, invites.
🤝
Co-parent
Verify achievements. Cannot distribute or configure.
🎒
Learner
View own data. Self-report. Cannot see siblings.
💝
Family
View progress. Send gifts. Nothing else.
📋
Advisor
Audit-read-only. For lawyers and finance.
Get Started
Three steps. Thirty seconds.

No install. No API keys. No developer console. Just a conversation.

1

Get Claude (free)

Download Claude for iPhone or Android. Create a free account. That's it.

2

Connect AllowanceAgent

Open claude.ai in your browser → Settings → Connectors → Add custom connector. Paste this URL:

https://allowme.dev/mcp

3

Start talking

Open Claude → tap + → Connectors → enable AllowanceAgent. Then say:
"Set up allowance for my family" or "Set up AI allowance for my team."

Connect
Your MCP server URL

Copy this URL and paste it in Claude → Settings → Connectors → Add custom connector.

https://allowme.dev/mcp

Free Claude plan: 1 custom connector. Works on mobile, desktop, and web.

Have an invite code? Connect first, then tell Claude your code.