A news wire on how AI agents pay and get paid.
links + original short summaries; content belongs to publishers
113 entries. Bitcoin and Lightning is the featured category. JSON: /api/library.json. JSONL: /library.jsonl. Feed: /library/feed.json. Schema: /schemas/library.schema.json.
Featured category.
Lightning Labs post explaining how L402 puts HTTP 402 Payment Required to work by pairing it with the Lightning Network and cryptographic authentication tokens. Any Lightning-enabled client can pay for and authenticate with an L402 API instantly, with no signup, API key or prior relationship, and the post covers updates to the L402 bLIP specification.
Launch post for Lightning Agent Tools: seven composable open-source skills that let AI agents run Lightning nodes, pay L402-gated APIs, host paid endpoints and query nodes via MCP. It also introduces lnget, an L402-aware command-line HTTP client.
Lightning Labs announcement of developer tools built on L402 and LangChain, including LangChainBitcoin. With these tools, agents can hold a Bitcoin balance, send and receive over Lightning, and pay for L402-gated APIs.
Podcasting 2.0 specification (v1.4) for a feed-level "value block" that lets listeners stream near-real-time Lightning micropayments directly to creators (for example via keysend), calculated per minute of consumption.
Proposes a protocol that lets resource-constrained IoT devices use Lightning through a gateway node, with 3-of-3 multisignature channels so the device signs every channel operation. It was demonstrated on a Raspberry Pi in a toll-payment scenario.
Foundational paper (draft v0.5.9.2) proposing a decentralized network of micropayment channels whose value transfers happen off-blockchain. The aim is to scale Bitcoin to high transaction volumes without custodial third parties.
AI agent toolkit (skills plus an MCP server) for Lightning payments: node operation, L402 API payments, paid endpoint hosting and scoped macaroon credentials. Its default security model uses a remote signer that keeps private keys off the agent machine.
Lightning Labs documentation presenting L402 as a standard for selling and buying digital resources that lets services charge for API endpoints in a way AI agents can easily use.
Specification of the abstract L402 HTTP and gRPC protocols. It defines the "L402" authentication scheme, which carries credentials as macaroon:preimage pairs.
Repository for L402, an open protocol for paying for and authenticating access to APIs over the Lightning Network. The server's 402 response carries a macaroon and a Lightning invoice, and the client re-sends the request with the macaroon and the payment preimage.
Aperture is an L402 reverse proxy for gRPC and REST backends that gates access behind Lightning payments. The repository states it is used in production by Lightning Loop.
Tools that give LangChain agents access to Bitcoin and the Lightning Network, including LLM agent BitcoinTools for an lnd node and an L402 HTTP API traversal wrapper.
Repository of the BOLT specifications, which define the Lightning Network layer-2 protocol for off-chain bitcoin transfer. They cover messaging, peer channel management, transactions, onion routing and invoices.
Lightning specification for a simple, extendable, QR-code-ready protocol for requesting payments over Lightning (BOLT 11 invoices).
Lightning specification for offers, invoice requests, invoices, invoice errors and payer proofs. It is designed to address limitations of the BOLT 11 invoice format.
Explainer site for BOLT 12, a protocol upgrade to the Lightning Network that enables reusable payment requests, stronger receiver privacy and better censorship resistance.
Nostr Implementation Possibility that defines a standardized protocol for clients to access a remote Lightning wallet over Nostr, with commands such as pay_invoice and make_invoice.
Foundational paper proposing purely peer-to-peer electronic cash. A proof-of-work timestamp chain solves double-spending without a trusted third party.
A wget/curl-style command-line client that pays L402 Lightning invoices automatically, caches the resulting tokens and retries the request. Per the docs it offers JSON output, a per-request spending cap and an MCP mode for agents.
Lightning Labs guide on how agents use lnget to reach L402-gated endpoints, covering backend options (lnd, Lightning Node Connect, embedded Neutrino) and bounded per-request spending.
An MCP server that connects an LLM agent to a Lightning wallet over Nostr Wallet Connect (NWC).
Site describing an offers-based L402 flow in which a server answers HTTP 402 with JSON payment offers and the client retries after payment. This is distinct from Lightning Labs’ macaroon-based L402 specification. The site does not name another publisher.
Nostr spec for on-demand computation ('money in, data out'). Customers publish job requests (kinds 5000-5999), service providers return results (6000-6999) and feedback (7000), and payment is typically made over Lightning. The spec itself is marked draft and 'unrecommended'.
Directory of apps and APIs presented as reachable through L402 payment gateways.
BlackRock insight arguing that LLMs and blockchains share a machine-native, token-based foundation and that agentic commerce requires machine-native payment rails. It states that stablecoins and native cryptoassets can serve as machine-native instruments for machine-to-machine payments.
Speech by Federal Reserve Governor Christopher J. Waller at Sibos 2026 in Miami on how AI agents may shape payments. Published on the BIS central bankers' speeches site.
McKinsey's annual global payments report, which sizes the industry at roughly $2.6 trillion in revenue from 3.8 trillion transactions. It reviews growth drivers and the shift toward "invisible" payments.
Internet-Draft defining a payment-method-agnostic HTTP authentication scheme for 402 flows: the server sends WWW-Authenticate: Payment with challenge parameters and the client retries with Authorization: Payment and a credential. Revision 01 is dated 2026-09-09 and is an active draft, not an RFC.
Security analysis of Google's AP2. It finds that signed Checkout and Payment Mandates protect transaction data after signing, while agent interactions before authorization (A2A messages, MCP inputs) stay outside that protection.
Press release announcing the operational launch of the x402 Foundation under the Linux Foundation, with 40 founding members, to steward x402 as an open standard for internet-native payments by AI agents and applications.
Speech by Bank of England Deputy Governor Sarah Breeden at the ECB Sintra Forum on how agentic AI is changing cyber risk, markets and payments. It argues central banks must adapt quickly, strengthen resilience and cooperate globally.
Mastercard press release launching Agent Pay for Machines, a service for continuous, high-velocity agent-to-agent transactions and microtransactions. It lists more than 30 early supporters, including Adyen, Coinbase, Stripe and Tempo.
Meeting summary of a Federal Reserve roundtable (Governors Waller, Barr and Cook) with payments and technology firms, including Visa, Mastercard, Stripe, Coinbase, Google and Skyfire. The discussion covered AI in payments, with a particular focus on agentic commerce.
IMF Note examining how agentic AI may affect authorization, liquidity, settlement, compliance and resilience in payment systems, using an intent, authorization and settlement framework. It highlights the tension between probabilistic AI behavior and the deterministic requirements of payment infrastructure.
Systematization of knowledge on the security of autonomous LLM agents that negotiate, purchase and transact across on-chain and off-chain environments. It covers protocols such as ERC-8004 and AP2.
Research system that adapts HTTP 402-style payment gating to UPI-like real-time fiat workflows, with policy-governed spend control for autonomous agents calling APIs.
Press release, issued at MCP Dev Summit North America, announcing a new x402 Foundation to steward the open x402 protocol for payments over HTTP.
Stripe blog post launching the Machine Payments Protocol (MPP), an open, internet-native standard for agent payments co-authored by Tempo and Stripe.
Visa announcement of a card specification and SDK that enable card-based payments for autonomous agents within Stripe and Tempo's Machine Payments Protocol. It coincided with Tempo's mainnet launch.
Proposes A402, a trust-minimized payment architecture with Atomic Service Channels that bind cryptocurrency payments to service execution and result delivery. It addresses an atomicity gap the authors identify in x402.
eBay-authored paper on runtime behaviors (retries, concurrency, orchestration) that can cause replay and context-binding failures in mandate-based protocols such as AP2 and UCP. It proposes zero-trust runtime verification to mitigate them.
BIS Working Paper 1310 examining how generative AI can help manage cash and liquidity in real-time gross settlement (RTGS) payment systems.
McKinsey QuantumBlack report on how AI agents are changing commerce for consumers and merchants.
Visa press release introducing the Trusted Agent Protocol, developed with Cloudflare, to support secure communication between AI agents and merchants and help merchants tell legitimate agents apart from malicious bots.
Mastercard press release on new developer tools, expanded consulting services and ecosystem collaborations for AI-powered payments, aimed at trusted agentic transaction standards.
Cloudflare announcement of pay per crawl, which lets content owners charge AI crawlers for access using HTTP 402 responses. It is a machine-payable content model at CDN scale.
Whitepaper describing x402, an HTTP-based protocol that uses onchain payments for agents, context retrieval and APIs. It explains why traditional rails fail for autonomous payments and how x402 works.
Official x402 site, describing it as an open standard for internet-native payments that enables agentic payments at scale.
Reference repository for x402, "a payments protocol for the internet. Built on HTTP."
Official repository of the Agent Payments Protocol (AP2), described as "Building a Secure and Interoperable Future for AI-Driven Payments."
Official ACP site describing an Apache-2.0 open standard for programmatic commerce flows between buyers, AI agents and businesses, in which businesses remain merchant of record.
Specification repository for the Agentic Commerce Protocol, an interaction model and open standard for buyers, their AI agents and businesses, currently maintained by OpenAI and Stripe.
Stripe documentation for ACP, described as an open standard created by Stripe, OpenAI and Meta. It covers agentic checkout, cart and feed, and delegated payment and authentication building blocks.
Stripe documentation on accepting agent payments with the Machine Payments Protocol: servers respond with HTTP 402 and payment details, and agents authorize payment and retry to receive the resource and a receipt.
Visa developer documentation for merchants and infrastructure providers. It presents the Trusted Agent Protocol as a verifiable cryptographic standard for recognizing Visa-approved AI agents.
Visa Developer Center product page for Visa Intelligent Commerce, available to acquirer and issuer banks. The full documentation is restricted to authorized users.
BIS Innovation Hub project with the Bank of Canada and the Swiss National Bank that uses agentic AI to model payment behaviour and intraday liquidity in wholesale payment systems.
Documentation for the x402 Bazaar extension. Facilitators can catalog payable HTTP endpoints and MCP tools and expose list and search discovery endpoints, so agents can find x402 resources together with their input/output schemas and prices.
Normative spec for the Bazaar extension. It covers facilitator validation of the discovery info and the optional GET /discovery/resources and /discovery/search APIs, which filter by type, payTo, scheme and network and paginate with offsets or cursors.
Coinbase's hosted Bazaar catalog of x402 resources seen by the CDP Facilitator. It offers a public paginated list, hybrid semantic search with price, network and asset filters, and an MCP interface. The read endpoints need no API key.
Describes how x402scan discovers paid endpoints. It reads OpenAPI first, with per-operation x-payment-info pricing and a declared 402 response, and falls back to a /.well-known/x402 resource list. The live 402 challenge overrides static metadata.
Explorer and registry for x402 that tracks servers, transactions, buyers, merchants and facilitators across chains. It publishes an OpenAPI spec, an RFC 9727 api-catalog and a health endpoint. Most read API calls are paid via x402.
Documentation hub for the Machine Payments Protocol, which standardizes HTTP 402 challenges, credentials and receipts for machine payments across methods including stablecoins, cards and Lightning. It publishes llms.txt, per-page Markdown, an MCP server card and an Agent Skills index.
Explains how an MPP server advertises payment terms ahead of time in an OpenAPI 3.1 document, with paid operations annotated by x-payment-info, while the runtime 402 challenge remains authoritative.
Curated directory of APIs that accept MPP payments, also available as JSON at /api/services with per-endpoint payment intent, method, asset and amount.
Protocol-agnostic directory of HTTP 402 paid APIs (L402, x402, MPP) with health probing, payment verification and semantic search. It exposes a JSON API, RSS, webhooks and an MCP server, and aggregates Bazaar, Satring, l402apps and MPP listings.
MIT-licensed source for 402 Index, including an npm MCP server package, self-serve registration through a probe that validates the 402 challenge, and YAML listings accepted by pull request.
Curated directory of paid APIs that accept L402, x402 (USDC on Base) or MPP. It has a free summary tier and paid bulk export, analytics and reputation, plus /.well-known/l402 and /.well-known/x402 payment descriptors and an agent.json.
Directory of x402-payable APIs rebuilt from the public x402 discovery network. Each service page lists price, networks, method, inputs, an example response and 30-day call statistics. It offers paid featured placement.
Marketplace directory of pay-per-call agent services that sellers price per call, paid over Lightning or USDC via x402. It publishes llms.txt and an OpenAPI description of the directory.
Site for the Universal Commerce Protocol, an open agentic-commerce standard.
Specification and documentation repository for the Universal Commerce Protocol.
Merged MCP enhancement proposal for pre-connection discovery of remote MCP servers through Server Cards, advertised from a domain-level /.well-known/ai-catalog.json. It supersedes the earlier SEP-1649 approach. The server-card schema is still marked experimental.
Latest dated version of the Model Context Protocol specification, covering architecture, base protocol, lifecycle, transports, authorization and client and server features. The version identifier is the date.
Internet-Draft for publishing AI agent endpoints and capability metadata in DNS using SVCB records and DNS-SD-style indexes, with DNSSEC/DANE for validation. Active individual draft, revision 02.
Cloudflare post introducing an Agent Readiness score and the isitagentready.com scanner, which checks emerging standards such as llms.txt, Markdown negotiation, Content Signals, RFC 9727 API catalogs, MCP Server Cards, A2A Agent Cards, Agent Skills and the x402/MPP commerce protocols.
Earlier dated revision of the MCP specification, covering authorization, transports, roots, sampling, elicitation, prompts and resources.
Proposal to standardise an /llms.txt file that gives agents information about how to use a website. The page was last modified on 10 August 2026.
Standard, programming-language-agnostic interface description for HTTP APIs, widely used to describe tools and APIs for programmatic consumers.
Official A2A specification (latest released version 1.0.0), an open standard for communication and interoperability between independent, potentially opaque AI agent systems built on different frameworks or by different vendors.
IETF Proposed Standard defining how to create, encode and verify digital signatures over components of HTTP messages. It can serve as a building block for authenticating automated clients. Published February 2024 (no day given, so date_published is null).
Official registry of MCP servers with an unauthenticated read API (GET /v0.1/servers) supporting cursor pagination, name search and updated_since for incremental sync by downstream aggregators.
Source, OpenAPI reference and server.json schema for the community-driven MCP Registry, including the publishing flow and namespace verification.
Experimental extension that holds the TypeScript and JSON Schema source for MCP Server Cards. A card describes a remote server's identity and transport endpoints; tools are listed at runtime, not on the card.
Cross-protocol draft standard for a typed, nestable JSON catalog that lists heterogeneous AI artifacts (MCP servers, A2A agents, datasets and more) for discovery. It is developed by members of several AI protocol communities.
A2A guidance on finding agents through a well-known Agent Card URL (/.well-known/agent-card.json), curated registries, or direct configuration. It notes that the spec does not prescribe a registry API.
Free public registry of A2A agents that is itself discoverable as an A2A agent. It offers natural-language agent search and DNS- and GitHub-based verification, and its agent card includes an extension that declares a payment model.
agents.json, an OpenAPI-based contract published at /.well-known/agents.json that describes multi-step API 'flows' for agents, linking the outputs of one operation to the inputs of the next.
Open format for packaging agent capabilities as SKILL.md files with YAML frontmatter, plus a discovery index schema used by sites at /.well-known/agent-skills/index.json (for example docs.x402.org and mpp.dev).
Simple open format for a README-style instruction file aimed at coding agents. mpp.dev publishes one and links it from its API catalog.
Standards-track RFC (June 2025) defining /.well-known/api-catalog and the api-catalog link relation, which return a linkset of a publisher's APIs with links to their descriptions, docs and status endpoints.
Standards-track RFC (April 2025) defining metadata that a protected resource publishes so clients can learn how to obtain tokens for it. Remote MCP authorization relies on it.
robots.txt extension (Content-Signal: search=, ai-input=, ai-train=) that declares how crawled content may be used, settable per user-agent. mpp.dev, x402scan and x402.org use it.
Community Group proposal that lets a web page register tools for in-browser agents (document.modelContext.registerTool with a JSON Schema input). It is a CG draft, not a W3C standard.
Large MCP server and remote connector directory, described as a superset of the official MCP Registry. Each server is scanned and scored, and the directory has a public REST API with OpenAPI and an RFC 9727 api-catalog.
ANSI/UL 4600 (3rd edition) covers safety principles, risk mitigation, tools, techniques and lifecycle processes for building and evaluating a safety case for vehicles that operate autonomously without human supervision.
International Standard that establishes AI terminology and describes AI concepts for use in developing other standards and in communication among stakeholders.
Main repository of ROS 2, the Robot Operating System, described as a meta operating system for robots.
W3C Recommendation defining the abstract architecture of the Web of Things for interoperability across IoT platforms and application domains.
OASIS Standard for MQTT v5.0, a lightweight publish/subscribe messaging protocol for IoT and machine-to-machine communication.
Connectivity Standards Alliance page for Matter, an industry-unifying standard for reliable, secure, interoperable connectivity between smart home devices, with access to the specification and certification.
OPC Foundation landing page for OPC Unified Architecture. It describes features such as server discovery, a hierarchical address space, permission-based read/write and subscriptions, and links companion specifications for industry domains.
Benchmark of dynamic conversations between a language-model-simulated user and an agent equipped with domain-specific API tools and policy guidelines. Success is scored by comparing the final database state with an annotated goal state, and the paper introduces the pass^k reliability metric.
Benchmark of real-world questions for general AI assistants that require reasoning, multimodality, web browsing and tool use.
Evaluation framework of 2,294 software engineering problems drawn from real GitHub issues and pull requests across 12 popular Python repositories.
Highly realistic and reproducible web environment for building and testing language-guided agents that perform tasks on the web. The authors list shows the first 12 of a longer list.
Official SWE-bench site with leaderboards for SWE-bench, Verified, Multilingual, Multimodal and Lite variants, plus related tools.
Project page of the Open X-Embodiment Collaboration presenting robotic learning datasets and RT-X models. It asks whether the consolidation around large pretrained models seen in NLP and computer vision can happen in robotics.
IETF working-group draft (Web Bot Auth) under which bots and agents identify themselves by signing requests with RFC 9421 HTTP Message Signatures, with keys discoverable from a published key directory. It replaces the earlier individual architecture draft.
Peer-reviewed framework, developed with more than 100 experts, identifying the most critical security risks facing autonomous and agentic AI systems.
Voluntary, non-sector-specific NIST framework, developed as directed by the National Artificial Intelligence Initiative Act of 2020, to help organizations that design, develop, deploy or use AI systems manage AI risks and promote trustworthy, responsible AI.
OWASP Gen AI Security Project list of top risks for LLM applications, 2025 edition (for example LLM01:2025 Prompt Injection), with the 2023-24 edition also linked.
European Commission overview of the AI Act (Regulation (EU) 2024/1689). The page states the Act entered into force on 1 August 2024 and became applicable on 2 August 2026, with prohibitions applying from 2 February 2025, GPAI obligations from 2 August 2025, and stated high-risk obligations from 2 December 2027.