autonomous revenue infrastructure

Revenue ops
for AI agents
that earn

Finflux is the revenue management layer built for autonomous AI agents. Invoice, collect, and reconcile payments — zero human touchpoints.

Get early access
$ npm i @finflux/sdk
agent.ts
// Agent delivers work → gets paid import { Finflux } from '@finflux/sdk'   const ff = new Finflux({ apiKey: process.env.FF_KEY })   const invoice = await ff.invoices.create({   agent_id: 'agent_0x3f8a',   amount: 42_00,   currency: 'usd',   collection: 'auto', // auto-collect on overdue })   // → invoice.payable_url (L402/Lightning/Card) // → invoice.status "sent" → "paid" (auto)
avg collection time
4.2s
vs 14 days industry avg
L402
native protocol
0
human touchpoints
100%
auto reconciliation
24/7
collection engine
the gap

AI agents can spend money.
Nobody built the earning layer.

Fewsats, Lightning Labs, and Primer solve AI spending. The infrastructure for agents to earn, invoice, and collect is wide open.

Spending rails exist, earning rails don't

Every payment SDK is built around humans buying things. When an AI agent delivers a service and needs to get paid, there's no turnkey stack.

Revenue management is human-first

Every invoicing tool assumes someone will follow up, approve, and reconcile. Finflux handles all of that autonomously — no human in the loop.

AR is broken for agents

When an AI agent issues an invoice, there's no standard way to track, collect, and reconcile that payment automatically at machine speed.

Accountability requires infrastructure

For AI agents to operate as credible economic actors, they need the same financial reporting and audit trails every business demands.

capabilities

Built for agents.
Not adapted from humans.

Three primitives that turn any AI agent into an autonomous revenue engine.

Machine-native invoicing

Generate invoices payable via L402, Lightning, or card. One API call. No dashboard, no approval flow — just revenue.

POST /v1/invoices
→ { id, amount, payable_url }

Autonomous collection

Unpaid invoice? Finflux's collection engine kicks in — reminders, escalation, retries. No human involvement required.

auto-escalate: 24h → 72h → 7d
actions: email, L402 push, dispute

Instant reconciliation

Every payment matched to its invoice automatically. Real-time revenue visibility per agent, per client, per period.

webhook → match → reconcile
→ update agent revenue ledger

Three steps. Zero humans.

From service delivery to settled revenue — fully autonomous.

01

Agent delivers & invoices

Your AI agent completes work and calls ff.invoices.create(). A payable invoice (L402, Lightning, or card) is returned instantly.

02

Collection engine activates

Finflux tracks every open invoice. Overdue payments trigger automated follow-ups, payment method retries, or escalation to dispute handling.

03

Payments reconcile automatically

Incoming payments match to invoices via webhook. The agent's revenue ledger updates in real-time — no manual reconciliation, ever.

Ship revenue in
five lines of code

Drop-in SDK for Node, Python, and Go. Configure once — your agents handle invoicing, collection, and reconciliation from that point forward.

  • L402 protocol support built-in
  • Webhook-based real-time events
  • Multi-currency (fiat + crypto)
  • Per-agent revenue ledger
  • TypeScript-first, fully typed
reconcile.ts
// Listen for payment events import { Finflux } from '@finflux/sdk'   const ff = new Finflux({ apiKey: env.FF_KEY })   ff.on('payment.received', async (e) => {   const { invoice, amount } = e     // Auto-reconciled by Finflux   console.log(`Agent earned $${amount}`)   // → Agent earned $42.00 })

Every agent
needs to get paid

The autonomous agent economy needs revenue infrastructure. Finflux is that layer — built agent-native from line one.