UPI Autopay and the quiet API-first payment revolution in Indian SMB SaaS

When an Indian SMB SaaS founder asks us about billing, the conversation used to start with "Razorpay or Stripe?" In 2026, it starts with "UPI Autopay or NACH?" The payment landscape for recurring B2B collections in India has changed faster than most founders realise, and the API surface has matured enough that even a two-person team can build production-grade billing.

This is the state of UPI-based recurring payments for Indian SMB SaaS as of mid-2026, written from the perspective of someone who has wired up billing for a textile ERP and watched the mandates flow.

What UPI Autopay actually is

UPI Autopay is NPCI's recurring payment mandate system, launched in 2020 and refined steadily. The user authorises a recurring debit from their bank account via UPI, with a defined amount, frequency, and expiry. The mandate is registered with NPCI, not with the merchant. The merchant initiates the debit, NPCI validates it against the mandate, and the bank executes it.

The technical flow:

1. Mandate creation. Your app calls Razorpay/PhonePe/Cashfree's mandate-create API, specifying the customer's UPI ID (vpa), amount, frequency, start date, and expiry. The customer receives a UPI collect request on their app. They approve it once. The mandate is now live.

2. Recurring debit. Each billing cycle, you call the debit API with the mandate ID and amount. The funds move. No customer interaction. No OTP. No 2FA.

3. Mandate management. The customer can view, pause, or cancel the mandate from any UPI app. Your webhook receives status updates.

This is not a credit card token stored in a vault. It is a recurring debit permission registered with the national payments infrastructure. The revocation path is the same UPI app the customer uses every day.

Why it matters for SMB SaaS

The old way of collecting recurring payments from Indian businesses was painful:

• Cheques that bounced, got lost, or required someone to physically visit a bank branch.
• NACH mandates that took 7-10 days to set up, required physical forms, and failed silently when bank details changed.
• Net banking standing instructions that worked until the bank migrated its platform and the instruction disappeared.
• Manual bank transfers where the customer forgot, you had to follow up, and the relationship strained over ₹2,000.

UPI Autopay replaces all of that with an API call and a one-time approval on a phone. For an SMB SaaS charging ₹5,000-25,000/month, the difference between "cheque-based collections" and "Autopay" is the difference between a billing team and a billing script.

The API surface, honestly assessed

Three providers dominate the UPI Autopay API market for SaaS companies in 2026:

Razorpay has the most mature Autopay integration. The API is well-documented, the dashboard is usable by non-technical finance teams, and the webhook reliability is solid. Pricing is 0-2% per transaction depending on volume. The tradeoff: Razorpay requires a full merchant account, which means KYC, bank account verification, and 5-7 business days of onboarding. For a registered Pvt Ltd with GST, this is straightforward. For a sole proprietor just starting out, it can be a hurdle.

Cashfree is the close second. Their Autopay API (they call it "Subscriptions") is cleaner than Razorpay's — fewer edge cases, better error messages, faster mandate creation. Their PG is also slightly cheaper at higher volumes. The dashboard is less polished but the API is more developer-friendly. We use Cashfree for Paraslace's billing.

PhonePe PG is newer to the merchant PG space but has the advantage of being integrated directly with the PhonePe consumer app, which is the dominant UPI app in India by transaction count. The mandate creation flow feels native to PhonePe users. The API is still maturing — we encountered a race condition in mandate status sync in March 2026 that their support team fixed in three days. Growing pains, but the distribution advantage is real.

// Cashfree — creating a UPI Autopay subscription (simplified)
const subscription = await cashfree.subscriptions.create({
  subscriptionId: `sub_${customerId}_${Date.now()}`,
  customerEmail: "owner@textileco.in",
  customerPhone: "919876543210",
  planId: "plan_pro_monthly",
  authAmount: 1, // ₹1 authorisation charge
  paymentMethod: {
    upi: { channel: "collect" },
  },
  subscriptionExpiryTime: "2027-06-01T00:00:00+05:30",
  firstChargeDate: "2026-07-01",
  returnUrl: "https://paraslace.in/billing/confirm",
});

The billing architecture we use

For Paraslace and future vertical products, the billing stack looks like this:

Paraslace (Next.js) → Billing Service (REST API)
                         │
                         ├─ Cashfree (Autopay mandates + UPI collect)
                         ├─ PostgreSQL (subscriptions, invoices, payments)
                         └─ Tally integration (GST invoicing, reconciliation)

The billing service is a thin Node.js API that wraps Cashfree's subscription API, maintains a local subscription state machine, and generates GST-compliant invoices on each successful debit. The invoices flow to Tally via the same pipeline we described in our GST automation post.

The state machine has five states: trial → active → grace_period (payment failed, retrying) → paused (customer paused mandate) → cancelled. Every transition emits a webhook to the Paraslace backend, which updates the customer's access accordingly.

What still breaks

Credit where it's due: UPI Autopay is the most reliable recurring payment mechanism in India today. But things still break.

Mandate expiry and renewal. Autopay mandates have a maximum validity period defined by NPCI (currently 1 year for most categories). The renewal flow is not a seamless extension — it requires a fresh mandate creation, which means a fresh UPI approval from the customer. Most customers approve it. Some don't. The churn from failed mandate renewal is real and underreported.

Bank-side failures. The mandate is valid. The debit request is authorised. The bank debits the account. Then the bank returns the funds the next day with a cryptic failure code. We see this in roughly 0.5-1% of debit attempts. The failure is always on the bank side — insufficient balance, account frozen, KYC issue, or a bank system outage. Cashfree and Razorpay retry automatically. Some debit attempts succeed on retry 2 or 3. Some never succeed and we have to follow up manually.

UPI Lite confusion. UPI Lite (the offline wallet feature that shipped in 2024) does not support Autopay mandates. A customer who has migrated their daily transactions to UPI Lite may not realise their Autopay mandate debits from the underlying bank account, not the Lite wallet. This causes confusion when the wallet shows a balance but the Autopay debit fails because the bank account is empty.

The compounding effect

The real impact of UPI Autopay on SMB SaaS is not the collection efficiency — though a 95%+ success rate vs the 70-80% of cheque/NACH is meaningful. The real impact is that revenue becomes predictable.

When you know ₹10,00,000 in recurring debits will succeed on the first of the month, you can hire sooner, invest in product, and plan capacity with a confidence that manual collections never allowed. The difference between "we'll probably collect ₹7L this month" and "we collected ₹9.6L on the 1st" changes how you run the business.

For the textile ERP we ship, the Autopay billing flow took three weeks to build and has processed roughly 400 mandate debits in its first four months with two failures that required manual intervention. That is a 99.5% success rate on payments, with zero human involvement per cycle. The previous manual collections process consumed roughly 15 hours of founder time per month.

UPI Autopay is not a feature. It is the billing infrastructure that makes Indian SMB SaaS viable at scale. Wire it up early. The alternative is a collections spreadsheet and a tense phone call every month.