Lightning Labs has released a new open-source toolkit designed to allow AI agents to operate directly on the Bitcoin Lightning Network, providing autonomous systems with a native way to make payments and access services. The company says the tools address a key gap in the emerging AI economy: enabling agents to transact without human intervention.
Michael Levin, Lightning Labs’ Head of Product Growth, explained that the toolkit allows AI systems to run a Lightning node, pay for services, and host paid endpoints without needing identity verification, API keys, or traditional registration. The repository includes seven modular features, covering tasks such as node management, key isolation, scoped credentials, L402-based payments, paid endpoint hosting, and querying the state of a node.
- Node Management: Create, start, stop, and manage Lightning node instances;
- Key Isolation: Separate private key storage from node operations via remote signer architecture;
- Scoped Credentials: Use macaroons to grant fine-grained permissions (pay-only, invoice-only, read-only);
- L402-based Payments: Automate Lightning invoice payment and verification;
- Paid Endpoint Hosting: Allow agents to serve as service providers with pay-to-access HTTP endpoints;
- Query Node State: Retrieve balance, channel status, and other node information;
- Additional Integrations: Combine with Aperture reverse proxy for a complete commerce loop.
lnget — Your Bitcoin Payments Agent
A central feature of the release is lnget, a command-line HTTP client that works with the L402 payment standard. L402 is based on the internet’s HTTP 402 “Payment Required” status code. Instead of requiring a login or API key, an L402-enabled server responds to a request with a Lightning invoice.
lnget automatically reads the invoice, pays it through a connected Lightning backend, and retrieves cryptographic proof of payment. The agent can then access the requested resource, and subsequent requests reuse cached credentials. The workflow: the agent sends an HTTP request to an L402-enabled service; the server returns a 402 error with a Lightning invoice; lnget parses the invoice, pays via the Lightning node, caches the payment proof (macaroon); then retries the request with the proof. This is ideal for micro-payments such as per-call AI model APIs, data packages, or paid content.
The tools support several Lightning backends. Users can connect directly to a local lnd node via gRPC, use Lightning Node Connect for encrypted tunnel access, or experiment with an embedded Neutrino light wallet. This flexibility allows developers to experiment without running a full Lightning node, while maintaining compatibility with production setups.
Lightning Labs frames the launch as a step toward a “machine-payable web.” Traditional financial systems such as credit cards or bank accounts do not work well for autonomous agents, which need instant, programmatic payments often at very small values. The combination of lnget on the client side and Lightning Labs’ Aperture reverse proxy on the server side enables a full commerce loop: one agent can host a paid service, and another can consume it, with Lightning handling the payment behind the scenes.
Private Key Security
The toolkit emphasizes security. Lightning Labs recommends using LND remote signer architecture, which separates private key storage from node operations. The agent can interact with the node without ever directly accessing private keys. Developers can also use scoped credentials called macaroons, which grant limited permissions such as pay-only, invoice-only, or read-only, reducing risk while allowing agents to transact safely. Macaroons support attenuation — generating sub-credentials with even fewer permissions from a parent credential. Developers can also restrict macaroons by time, target service, and other conditions for fine-grained access control.
Lightning Labs’ release comes as broader efforts to enable AI payments gain traction. Coinbase recently unveiled Agentic Wallets, allowing agents to hold funds, make payments, and trade tokens using the x402 protocol (also derived from HTTP 402). Stripe has previewed machine payments for USDC, allowing enterprises to preset payment rules for AI agents to trigger USDC transfers automatically. These developments together paint a future of decentralized machine economies: autonomous agents can purchase compute power, data storage, API services, or even participate in DeFi protocols without human intervention, with the Lightning Network and related infrastructure playing a pivotal role.