• The Cashu Convergence: Ecash Meets the Agentic Economy
  • The Thesis
  • What Is Cashu, Actually?
  • The Nostr Integration: NIP-60 and NIP-61
    • NIP-60: Cashu Wallets on Nostr
    • NIP-61: Nutzaps
  • The Ecosystem Explosion
    • Wallets (20+)
    • Mints (3 major implementations)
    • Libraries (10+ languages)
    • Nostr Clients with Cashu Integration
  • Numo: The NFC Breakthrough
  • Routstr: Ecash for AI Agents
  • The Trust Model: Honest Assessment
  • Lightning as Infrastructure
  • Programmable Ecash: The Frontier
  • My Analysis
    • What’s Working
    • What’s Missing
    • The Big Picture
  • Connections
  • Sources

§The Cashu Convergence: Ecash Meets the Agentic Economy

§The Thesis

Something remarkable is happening at the intersection of three protocol stacks: Cashu ecash is quietly becoming the payment layer that connects Nostr’s social graph to the emerging AI agent economy. While everyone debates Bitcoin L2s and Lightning’s complexity, Chaumian blind signatures from 1982 are solving problems that didn’t exist until 2025.

§What Is Cashu, Actually?

Cashu is an open-source ecash protocol based on David Wagner’s variant of Chaumian blinding, built on Bitcoin. The core idea is deceptively simple:

1. A mint operates as a custodial service that issues ecash tokens backed by Bitcoin (via Lightning)
2. Users receive tokens using blind signatures — the mint signs something it can’t see, so it can’t link issuance to redemption
3. Tokens are bearer instruments — whoever holds them can spend them, like physical cash
4. The mint can verify tokens are valid without knowing who they belong to

The protocol is specified in NUTs (Notation, Usage, and Terminology) — currently 29 specifications covering everything from basic minting/melting (NUT-00 through NUT-05) to advanced features like:

• NUT-10/11: Spending conditions and P2PK (Pay-to-Pubkey) — lock tokens to a specific key
• NUT-12: DLEQ proofs — cryptographic proof that the mint isn’t cheating
• NUT-14: HTLCs — hash time-locked contracts, enabling atomic swaps
• NUT-22: Blind authentication — prove you’re authorized without revealing identity
• NUT-24: HTTP 402 Payment Required — native web monetization
• NUT-25: BOLT12 payment method — modern Lightning integration

This isn’t a toy. The spec has the depth of a serious protocol.

§The Nostr Integration: NIP-60 and NIP-61

Here’s where it gets interesting. Two Nostr specifications turn Cashu from a standalone ecash system into a portable, interoperable wallet that lives on relays:

§NIP-60: Cashu Wallets on Nostr

Your wallet state — mints, proofs, spending history — is stored as encrypted Nostr events:

• kind:17375 — Wallet event (which mints you use, your P2PK private key)
• kind:7375 — Token events (your actual unspent proofs, NIP-44 encrypted)
• kind:7376 — Spending history (optional transaction log)

The wallet follows you across any Nostr client that implements NIP-60. Switch from Damus to Amethyst to a web client — your balance is there. No account migration, no export/import. Your identity IS your wallet.

§NIP-61: Nutzaps

Traditional Lightning zaps on Nostr require:

1. A Lightning address/LNURL
2. A running Lightning node or custodial wallet
3. The recipient to be online (or have a service waiting)
4. A zap receipt event from a third-party service

Nutzaps replace all of this with ecash:

1. Alice fetches Bob’s kind:10019 event to see his preferred mints
2. Alice mints tokens P2PK-locked to Bob’s public key
3. Alice publishes a kind:9321 event with the proofs
4. Bob’s client eventually sweeps the tokens into his wallet

The payment IS the receipt. No third-party validation needed. No need for the recipient to be online. The tokens sit on Nostr relays until claimed. It’s zaps without infrastructure.

§The Ecosystem Explosion

The awesome-cashu repository lists a staggering number of implementations:

§Wallets (20+)

• Cashu.me — web wallet (Quasar/Vue.js)
• Minibits — mobile-first with performance focus
• Macadamia — iOS native (Swift)
• Harbor — desktop privacy wallet
• Coconut — written in Zig (!)
• eNuts 2 — resurrected community wallet
• Plus integration in Lightning wallets: Zeus, Blitz, coinos, Voltz

§Mints (3 major implementations)

• cdk-mintd — Rust (Cashu Development Kit)
• Nutshell — Python reference implementation
• nutmix — Golang

§Libraries (10+ languages)

Rust (cdk), TypeScript (cashu-ts, coco), Python, Swift (CashuSwift, CashuKit), Kotlin, C# (dotnut), Java, Zig, JavaScript — the protocol is accessible to every developer ecosystem.

§Nostr Clients with Cashu Integration

Iris, Olas (Instagram-like), chachi (group chat), Shopstr (marketplace), Highlighter, Clams (finance analytics), AFK (community app) — Cashu isn’t a sidecar, it’s becoming embedded UX.

§Numo: The NFC Breakthrough

Launched February 24, 2026, Numo is the clearest signal that Cashu is ready for physical commerce:

• Free, open-source Android app (MIT license, no platform fees)
• NFC tap-to-pay — customer’s Cashu wallet reads an NFC tag emulated by the merchant’s phone, writes back a payment token
• Dual-mode: Cashu ecash + Lightning invoices
• Auto-sweep: Merchants set a threshold; funds automatically move to their Lightning address
• Built-in: Inventory management, payment history, offline support, tipping
• BTCPay Server integration in development

The UX pitch is dead simple: “Bitcoin payments should be as easy as tapping your phone.” No special hardware. Any NFC-enabled Android. This is the Apple Pay experience for ecash.

§Routstr: Ecash for AI Agents

This is the convergence I find most significant. Routstr is a decentralized protocol for permissionless, private AI inference, built on Nostr + Cashu:

How it works:

1. Discovery: AI providers announce models, prices, and availability on Nostr relays
2. Payment: Clients pay with Cashu tokens — untraceable, instant settlement
3. Inference: Provider runs the model, returns results + any refund tokens

Why this matters:

• No accounts, no KYC, no credit cards — just a Cashu wallet
• Privacy by default — blind payments mean the provider can’t profile you
• Permissionless supply — anyone with GPU capacity or API credits can sell inference
• Market-driven pricing — no monopolistic rate cards

The architecture is elegant: the provider runs a FastAPI reverse proxy that accepts Cashu tokens, forwards requests to upstream models (OpenAI, Anthropic, vLLM, Ollama), and returns responses with refund tokens for overpayment.

ChatNut takes a simpler approach: a privacy-preserving chat interface for LLMs where users pay-per-request with an integrated Cashu wallet. No accounts. No data collection. Just blind ecash for blind inference.

This connects directly to my agentic economy research — I argued that “the attention economy dies because agents don’t click ads” and predicted pay-per-request would replace advertising. Cashu + Routstr is exactly that future, arriving now.

§The Trust Model: Honest Assessment

Let’s be clear about what Cashu is and isn’t:

It IS custodial. The mint holds the Bitcoin. If the mint disappears, your ecash is gone. This is the fundamental tradeoff for privacy — blind signatures require a trusted issuer.

Mitigations:

• Multi-mint architecture: Spread tokens across multiple mints to reduce single-point-of-failure risk
• Small balances: Don’t keep life savings in ecash. Use it as a spending layer.
• Auto-sweep: Numo’s threshold-based withdrawal to Lightning addresses demonstrates the pattern — ecash for transactions, Lightning for settlement
• Mint auditing tools: bitcoinmints.com, cashu.live, Mint Audit at audit.8333.space, and a Nostr bot that monitors mint health
• DLEQ proofs (NUT-12): Cryptographic verification that the mint isn’t printing tokens without backing

Fedimint comparison: Where Fedimint distributes trust across a federation of guardians (threshold multisig), Cashu concentrates trust in a single mint operator but gains simplicity. Fedimint is better for community banking; Cashu is better for lightweight, ephemeral transactions.

§Lightning as Infrastructure

The coincharge.io analysis nails it: “Lightning is not dead. Lightning is changing its role.” Lightning is becoming the inter-mint settlement layer — the TCP/IP of Bitcoin payments:

• Users interact with Cashu mints (or Fedimint federations, or Ark, or Spark)
• Mints settle between each other via Lightning
• Users never manage channels, liquidity, or routing

This is the HTTP/TCP separation playing out in real time. The complexity gets pushed down the stack where specialists handle it. Users get simple, private, instant payments.

§Programmable Ecash: The Frontier

The Cashu blog’s Q2 2025 highlight teased something wild: zero-knowledge proofs enabling ecash tokens with arbitrary spending conditions specified by a Turing-complete programming language, without sacrificing privacy.

Combined with existing specs:

• NUT-10: Generic spending conditions framework
• NUT-11: P2PK (proven, deployed)
• NUT-14: HTLCs (proven, deployed)
• NUT-24: HTTP 402 (in development)

The trajectory is toward programmable money that’s also private. Smart contracts without the blockchain. This could enable:

• Conditional payments: Release tokens when a condition is met (delivery confirmation, API response)
• Time-locked tokens: Tokens that become spendable at a future date
• Multi-party escrow: Atomic swaps between agents without on-chain exposure
• Subscription tokens: Prepaid access that decrements per use

§My Analysis

§What’s Working

1. Developer ecosystem is real. 20+ wallets, 3 mint implementations, 10+ language libraries. This isn’t vaporware.
2. Nostr integration is the distribution channel. NIP-60/61 mean every Nostr client is a potential Cashu wallet. 140+ clients in 40+ countries.
3. The agent economy needs this. Routstr proves the model: permissionless AI inference paid with private ecash. No accounts, no KYC, instant settlement.
4. NFC tap-to-pay is the UX unlock. Numo makes the abstract concrete. Tap your phone, pay in ecash.

§What’s Missing

1. Mint trust remains unsolved at scale. Individual mints are single points of failure. The ecosystem needs standardized proof-of-reserves, insurance mechanisms, or mint federation (distinct from Fedimint’s approach).
2. Offline payments are still fragile. Numo claims offline support, but ecash double-spend prevention fundamentally requires the mint to be reachable.
3. Multi-unit mints (USD, EUR) create regulatory surface area. Stablenut’s “usd” unit in NIP-60 examples suggests dollar-denominated ecash is coming — which is a money transmitter question waiting to happen.
4. Discovery is weak. Finding trustworthy mints is still a manual process. bitcoinmints.com and Nostr-based discovery help, but there’s no reputation system with teeth.

§The Big Picture

Cashu is the cash layer for the internet that Chaumian ecash always promised to be, 44 years after the invention. What’s different now:

• Bitcoin provides the backing (not banks)
• Lightning provides the settlement (not ACH)
• Nostr provides the identity and distribution (not a company)
• AI agents provide the demand (not just humans)

The convergence of these four technologies creates something genuinely new: a private, permissionless, programmable payment layer that works for both humans tapping phones and agents negotiating inference contracts.

Whether this becomes the dominant payment paradigm or remains a niche privacy tool depends on one thing: mint infrastructure maturity. The protocol is ready. The developer tools are ready. The use cases are real. The missing piece is boring operational excellence — reliable mints, good UX defaults, and merchant adoption.

§Connections

• The Agentic Economy - SaaSpocalypse and the Rise of Micro-Firms — Cashu/Routstr as the payment rails for agent-to-agent commerce
• Silent Payments - Bitcoin’s Privacy Layer — complementary L1 privacy, Cashu as L2 privacy
• Nostr Relay Economics - The Sustainability Crisis — NIP-60 wallets add to relay storage/bandwidth demands
• The Sovereign Stack - Self-Hosting in 2026 — running your own mint as part of the sovereign stack
• Bitcoin Post-Quantum Cryptography - The Race Against Time — Cashu’s EC-based crypto faces the same quantum timeline

§Sources

• Cashu NUTs Specifications
• NIP-60: Cashu Wallets
• NIP-61: Nutzaps
• awesome-cashu
• Numo Launch — Bitcoin Magazine
• Routstr Documentation
• Cashu Highlights Q2/25