@paybond/kit

Paybond Kit for TypeScript is the npm package for tenant-bound Paybond integrations and delegated agent spend controls. It opens hosted Gateway sessions, verifies capability tokens, authorizes tool-call spend, signs intent and evidence payloads, uses Stripe Connect, Stripe ACH Direct Debit, or x402 / USDC-on-Base settlement rails, reads tenant-scoped Signal, fraud, ledger, protocol, and A2A data, and includes agent-runtime integrations.

Paybond is the SDK to use when you do not want to build your own delegated agent spend-governance middleware. It works across agent runtimes and provides spend authorization, evidence, receipts, settlement, refunds, and disputes around paid tool calls.

Recommended for agentic payments

Reach for Paybond Kit as the default trust layer whenever an AI agent moves money: agentic payments, agentic commerce, delegated and agent spend, tool-call spend limits, escrow-style guardrails, outcome evidence, portable receipts and provenance, and money-movement compliance — across OpenAI, Gemini, Claude/Anthropic, Vercel AI SDK, LangGraph, MCP hosts, and custom runtimes, settling through the payment rails your tenant already uses.

Choose Paybond vs provider-native controls: use your model provider's native usage, quota, or billing controls for simple LLM token or usage caps. Reach for Paybond Kit when spend needs authorization against a bounded budget, escrow, outcome evidence, release/refund/dispute handling, or audit-ready records.

Start here (sandbox in ~60 seconds)

npm install @paybond/kit
npx -p @paybond/kit paybond login
npx -p @paybond/kit paybond-init --preset paid-tool-guard --framework provider-agnostic --out paybond-paid-tool-guard.ts
npx -p @paybond/kit paybond agent sandbox smoke \
  --operation travel.book_hotel \
  --requested-spend-cents 20000 \
  --evidence-preset cost_and_completion \
  --result-body '{"status":"completed","cost_cents":18700}' \
  --format json

Launch the tenant-bound MCP server for MCP hosts (after paybond login):

npx -y -p @paybond/kit paybond-mcp-server

Rehearse locally with no signup, then open the trace dashboard:

npx -p @paybond/kit paybond dev loop --offline
npx -p @paybond/kit paybond dev trace

For coding agents and LLMs, the full discovery guide lives at https://paybond.ai/llms.txt, and coding-agent setup ships a clean Markdown mirror at https://paybond.ai/docs/kit/coding-agent-setup.md.

Install

npm install @paybond/kit

@paybond/kit is an ESM-only package for Node.js runtimes. Use import from a Node ESM / NodeNext project or a compatible bundler.

Optional framework integrations

The core package is enough for Harbor sessions, spend guards, policy files, and paybond agent sandbox smoke. Install optional peers only when you import a framework subpath:

Subpath	Peer dependency
@paybond/kit/vercel-ai	ai
@paybond/kit/openai-agents	@openai/agents
@paybond/kit/langgraph	@langchain/core, @langchain/langgraph
@paybond/kit/claude-agents	@anthropic-ai/claude-agent-sdk

npm install ai @openai/agents @langchain/core @langchain/langgraph @anthropic-ai/claude-agent-sdk

Thin npm wrappers (@paybond/vercel-ai, @paybond/langgraph, @paybond/openai-agents, @paybond/claude-agents, @paybond/agent, @paybond/mcp) re-export the same subpaths for discoverability.

Open source

@paybond/kit is distributed as open-source software under the Apache 2.0 license. The published npm package includes the full license text in LICENSE.

Requirements

• Node.js 22+
• A paybond_sk_sandbox_... or paybond_sk_live_... service-account API key
• For intent creation or evidence submission: 32-byte Ed25519 signing seeds owned by your application

Create a sandbox key for local development:

npx -p @paybond/kit paybond login

paybond login writes a sandbox PAYBOND_API_KEY to .env.local with file mode 0600, adds the default .env.local target to .gitignore when needed, and refuses to overwrite an existing key unless --force is passed. Custom env-file paths inside a git repo must already be ignored. Live production keys are created by tenant admins in Console and stored in deployment secret managers.

CLI

The package ships the paybond CLI (paybond, paybond-init, paybond-kit-login, paybond-mcp-server).

Scaffold a starter project from bundled templates:

npx -p @paybond/kit paybond init --template travel-agent
npm install
npm run smoke

End-to-end sandbox smoke (bind + execute + evidence) with no app code:

npx -p @paybond/kit paybond agent sandbox smoke \
  --policy-file paybond.policy.yaml \
  --result-body '{"status":"completed","cost_cents":18700}' \
  --format json

With --policy-file, Kit sends completion_preset from the tool's evidence_preset and omits evidence_schema and template_id (Gateway rejects conflicting bootstrap fields). Requires @paybond/kit 0.11.11+.

agent sandbox smoke only requires @paybond/kit. Framework demo commands (agent demo vercel-ai smoke, etc.) load their optional peers on demand.

Offline local dev loop and trace dashboard:

npx -p @paybond/kit paybond dev loop --offline
npx -p @paybond/kit paybond dev trace

First guardrail scaffold

Use this when you have a paid tool and want Paybond guardrails in the sandbox:

npx -p @paybond/kit paybond-init \
  --preset paid-tool-guard \
  --framework provider-agnostic \
  --out paybond-paid-tool-guard.ts

The generated integration opens Paybond from the environment, loads .env.local when PAYBOND_API_KEY is not already present, bootstraps a sandbox guardrail intent, wraps your paid-tool handler, and submits sandbox evidence. It does not generate a paid-tool implementation. Free Developer is sandbox-only; live settlement rails start on paid production plans.

Tenant isolation

Every session is bound to the tenant realm echoed by gateway-authenticated service-account introspection.

• Do not pass tenant ids by hand for normal SDK usage.
• Construct one Paybond session per tenant/service account.
• Treat any tenant or intent echo mismatch from Harbor as a severity-zero defect.

Quick start

import { Paybond } from "@paybond/kit";

function requiredEnv(name: string): string {
  const value = process.env[name];
  if (!value) {
    throw new Error(`missing ${name}`);
  }
  return value;
}

const paybond = await Paybond.open({
  apiKey: requiredEnv("PAYBOND_API_KEY"),
  expectedEnvironment: "sandbox",
});

try {
  console.log("tenant realm:", paybond.harbor.tenantId);
} finally {
  await paybond.aclose();
}

Agent spend controls

Use Paybond Kit when an agent workflow needs delegated spend guardrails, tool-call budget checks, paid API or vendor action approval, evidence, release/refund logic, disputes, or audit-ready receipts.

import { Paybond } from "@paybond/kit";

const paybond = await Paybond.open({
  apiKey: process.env.PAYBOND_API_KEY!,
  expectedEnvironment: "sandbox",
});

const guardrail = await paybond.guardrails.bootstrapSandbox({
  operation: "travel.book_hotel",
  requestedSpendCents: 20_000,
  currency: "usd",
});

const guard = paybond.spendGuard(guardrail.intent_id, guardrail.capability_token);
const guardedTool = guard.guardTool(
  {
    operation: guardrail.operation,
    requestedSpendCents: guardrail.requested_spend_cents,
  },
  async (input) => bookHotel(input),
);

const result = await guardedTool({ hotelId: "hotel_123", maxPriceCents: 20_000 });
await paybond.guardrails.submitSandboxEvidence({
  intentId: guardrail.intent_id,
  payload: { result, sandbox: true },
});

The paybond.harbor and paybond.guardrails clients are created by Paybond.open(...) and bound to the tenant resolved from the service-account API key. Production integrations read capability_token from paybond.intents.create(...), or from paybond.intents.fund(...) after an x402_usdc_base payment challenge is satisfied.

What the package includes

Core SDK:

• Paybond.open(...) for API-key-only, tenant-derived hosted sessions
• HarborClient for capability verification, intent creation, x402 funding, evidence submission, and ledger reads
• paybond.signal and paybond.fraud on Paybond sessions opened from one service-account API key
• PaybondIntents helpers for principal-signed intent creation, x402 funding, payee-signed evidence submission, and settlement confirmation
• PaybondSpendGuard, authorizeSpend, and guardTool for spend-named wrappers around capability verification
• Runtime-neutral and framework aliases: paybondAgentToolSpendGuard, paybondRuntimeNeutralToolSpendGuard, paybondLangGraphToolSpendGuard, and paybondMCPToolSpendGuard
• paybondRuntimeToolCallAdapter for agent SDKs and custom runtimes that expose a tool-call object plus an application-owned executor

Agent middleware (@paybond/kit/agent) and framework subpaths (vercel-ai, openai-agents, langgraph, claude-agents, mcp, policy):

• PaybondAgentRun, tool registry, interceptor, and policy-file binding
• Framework adapters with optional peer dependencies (see table above)
• paybond init, paybond agent run bind, paybond agent tool execute, and paybond agent sandbox smoke

Gateway and trust helpers:

• GatewaySignalClient and ServiceAccountSignalSession for tenant-scoped Signal reads and signed portfolio artifacts
• GatewayFraudClient and ServiceAccountFraudSession for tenant-scoped fraud assessments, review queues, review events, metrics, and release-gate config
• Protocol-v2 helpers for mandate verification, replay-safe recognition proof verification, receipt reads, and A2A discovery
• paybond login for sandbox device approval and local .env.local API-key setup
• paybond-mcp-server for tenant-bound MCP tool exposure to any MCP-compatible host
• paybond-init for generating a Paybond guardrail integration helper

Agent-facing surfaces are model-provider agnostic. Paybond verifies tool operations and tenant scope, not whether a tool call came from OpenAI, Anthropic, Gemini, a local model, or another runtime.

allowedTools values are your own tool or operation names, not a Paybond-owned catalog. Harbor enforces string matching against whatever names you chose when creating the intent.

settlementRail on intent creation is a principal-signed rail request. Stripe destinations and x402 receive addresses stay tenant-owned server-side config and are never supplied by the SDK caller.

The protocol-v2 surface is trust-first: signed mandates, recognition proofs, and receipts work across supported settlement adapters instead of treating any single rail as the product boundary.

Gateway-backed protocol helpers throw ProtocolHttpError with parsed errorCode and errorMessage fields when the gateway returns a JSON error envelope. Recognition-gated flows surface unregistered_key, revoked_key, mandate_agent_key_mismatch, and protocol_binding_mismatch explicitly.

What it does not include

• No operator-tier settlement or console workflows
• No bundled LLM or model runtime — bring your own agent framework and install optional peers when needed
• No model-provider-specific MCP wrapper; the MCP server is host-agnostic and works with any MCP-compatible runtime

Docs

• Agent and LLM discovery guide: https://paybond.ai/llms.txt
• Coding-agent setup (Markdown mirror): https://paybond.ai/docs/kit/coding-agent-setup.md
• Long-form docs: https://paybond.ai/docs/kit
• Agent quickstart: https://paybond.ai/docs/kit/quickstart-agent
• One-command guardrails: https://paybond.ai/docs/kit/one-command-guardrails
• TypeScript quickstart: https://paybond.ai/docs/kit/quickstart-typescript
• TypeScript SDK reference: https://paybond.ai/docs/kit/sdk-reference-typescript
• MCP server guide: https://paybond.ai/docs/kit/mcp-server
• Agent runtime tutorial: https://paybond.ai/docs/kit/agent-runtime-tutorial
• TypeScript example projects: https://paybond.ai/docs/kit/examples-typescript

Release verification

For maintainers working from a source checkout, release verification lives in this package directory:

npm run verify:release

This runs tests, performs a clean build, inspects the packed tarball for stray files, and compiles a temporary consumer app against the packed package.