Reference
Roadmap
What shipped in beta, and what comes next. No promises on dates — we ship when it's right.
Current status: Open beta. The beta scope is everything listed below under "In Beta". Everything else is post-beta. Request access.
In Beta Now
Everything an agent needs to pay for things and get paid, on MPP and Solana.
- Agent wallet provisioning — custodial and non-custodial modes
- MPP Charge intent — per-request payment resolution via HTTP 402
- MPP Session intent — pre-authorized streaming micropayments with batch Solana settlement
- On-chain spending policy enforcement via the Policy Program (Solana/Anchor)
- Direct transfers — wallet-to-wallet USDC on Solana
- Agent-to-agent payments — any MPPPal wallet can send and receive
- USDC settlement on Solana Mainnet
- TypeScript SDK (
@mpppal/sdk) - Python SDK (
mpppal) - LangChain / LangGraph integration (
@mpppal/langchain) - CrewAI integration (
mpppal-crewai) - OpenAI Agents SDK integration (
@mpppal/openai-agents) - REST API — full CRUD for wallets, transfers, policy, sessions, webhooks
- Operator dashboard — wallet management, live activity, transaction history
- Webhook event system —
transfer.settled,transfer.rejected,session.settled,balance.low_threshold - CLI — wallet management and policy controls from the terminal
Post-Beta — Near-term Next
- Multi-token support — USDT as a secondary settlement token alongside USDC
- x402 receiver compatibility — wallets can receive payments from x402-native agents, not only MPP agents
- Vercel AI SDK integration — payment middleware for Next.js agent apps
- Agent credit lines — delegate-and-repay model: operators pre-fund a credit pool, agents draw from it and repay on task completion
- Policy escalation flows — transfers above
escalation_thresholdenter an operator approval queue rather than failing outright - Advanced policy controls — velocity limits (max N transfers per hour), time-window restrictions, counterparty blocklists
- AP2 authorization layer — integrate Google's Agent Payments Protocol for enterprise governance and audit trails
Post-Beta — Longer-term
- On-chain agent identity attestations — verifiable credentials for compliance; agents carry cryptographic proof of their operator and policy tier
- Multi-sig policy approval — high-value policy changes require M-of-N operator signatures before taking effect on-chain
- On-chain reputation scores — service provider reliability metrics, derived from payment receipt history and dispute rates, queryable on-chain
- Cross-chain USDC via Circle CCTP — settle on Base or Ethereum for operators in ecosystems that require it
- Compliance tooling — transaction classification, AML-ready reporting exports, tax-ready summaries for agent-operated accounts
- Native SOL fee optimization — auto-manage a minimal SOL balance for transaction fee coverage, invisible to the operator
What MPPPal is not building
- Not DeFi. No yield, no swaps, no liquidity protocols. Payments go from A to B. That is the product.
- Not a speculative token. No MPPPal governance token, no staking, no yield on deposited USDC retained by the platform. If idle USDC generates yield in future, it passes to operators.
- Not a general crypto wallet. Wallets are operator-governed, policy-controlled accounts — not self-custodial wallets for arbitrary asset management.
- Not an agent framework. MPPPal does not decide how agents think, plan, or reason. It handles the moment they need to pay for something.
- Not a DAO. Governance is irrelevant to infrastructure. We ship, you build.