High-Risk Payment Processing in Nepal: eSewa, Khalti and Bank Rails
By PayEurasia Team · 11 August 2026 · 11 min read
Last updated 11 August 2026
Nepal is a small, wallet-led market where provider choice and reconciliation discipline decide whether high-risk volume is sustainable. Here is how it works.
Nepal is the smallest of the four markets PayEurasia covers, and the one where merchants most often assume a regional integration will simply extend. It does not extend automatically: acceptance is wallet-led, provider depth is thinner than in India or Bangladesh, and underwriting for high-risk verticals is correspondingly more conservative. Treated on its own terms, it is a workable and under-served market.
What this guide covers
- Acceptance: wallets first, bank rails for larger tickets
- Why provider depth matters more here
- Underwriting expectations
- Reserves, limits and NPR settlement
- Reconciliation on a smaller book
- Risk controls
- Nepal as part of a regional footprint
- A pragmatic launch plan
Acceptance: wallets first, bank rails for larger tickets
eSewa and Khalti are the wallets with meaningful consumer adoption, and connectIPS plus direct bank transfer carry account-to-account flows. Cards play a minor role, and cross-border card acceptance is not a viable primary rail for a foreign merchant collecting from Nepali customers.
Design the checkout to lead with wallets, offer bank transfer for larger amounts, and keep the payment reference visible throughout. In smaller markets the customer's tolerance for an unfamiliar flow is lower, not higher.
Why provider depth matters more here
In a market with fewer providers, a single suspension or maintenance window represents a larger share of available capacity. Redundancy is therefore more valuable in Nepal than in India, not less, even though there are fewer routes to choose from.
Configure at least two routes where the methods allow it, and make failover automatic. Where only one route exists for a method, make sure your checkout can gracefully steer customers to an alternative method rather than failing outright.
Underwriting expectations
Expect the standard pack — entity registration, ownership to ultimate beneficial owners, director identification, live product URLs, published refund terms, settlement account in the entity's name — plus a clear description of the funds flow and realistic volume expectations for a market of this size.
An inflated forecast is more damaging in Nepal than elsewhere, because the discrepancy between forecast and reality is immediately obvious against known market volumes.
Reserves, limits and NPR settlement
Rolling reserves and conservative launch limits apply as they do regionally. Collections are NPR-denominated; the conversion point, rate source and timing should be fixed in writing if funds leave NPR.
Because absolute volumes are smaller, fixed per-transaction fees weigh more heavily on unit economics than percentage rates. Compare providers on total cost at your actual ticket size, not on the headline rate.
Reconciliation on a smaller book
Small books make reconciliation easy to neglect and easy to fix. Treat the signed webhook as the source of truth, keep handlers idempotent, and reconcile initiated, confirmed, reversed, refunded and settled counts daily.
Establish the discipline while volume is low. Merchants who scale a broken reconciliation process into a larger market discover the gaps at the worst possible moment.
Risk controls
Cap first deposits, match withdrawal destinations to funding identities, run velocity checks, and review any account that funds from multiple wallets in quick succession. In a smaller market a handful of abusive accounts can meaningfully move your dispute ratio.
Document the controls and enforce them consistently. Written, enforced controls are the evidence that supports a limit increase at review.
Nepal as part of a regional footprint
Most merchants reach Nepal after Bangladesh or India. Running it on the same API, the same webhook contract and the same consolidated settlement view keeps the marginal operating cost low enough to justify a smaller market. The Nepal gateway guide covers provider selection.
Enable it as a configuration change on an existing integration rather than as a separate project, and the market pays for itself far sooner.
A pragmatic launch plan
Enable wallets first with low limits, add bank transfer once wallet reconciliation is clean, ramp in agreed steps, and review limits and reserve at the agreed trigger. Keep refunds fast and visible throughout.
Expect the ramp to be measured in customers rather than in orders of magnitude. Steady, accurate growth is what unlocks better terms here.
Frequently asked questions
Can foreign merchants collect NPR payments?
Yes, through providers holding local acceptance relationships, subject to underwriting and a documented settlement structure.
Are eSewa and Khalti available to high-risk merchants?
It is decided per merchant by vertical, provider and underwriting outcome rather than by category alone.
Is Nepal worth enabling for a small volume?
Usually yes, if it runs on an existing regional integration; the marginal engineering and reconciliation cost is what determines viability.
What is connectIPS?
A Nepali interbank payment service used for account-to-account transfers, complementing wallet acceptance for larger tickets.
How should I compare provider pricing?
At your actual ticket size, including fixed per-transaction fees, payout fees and any reserve cost, not on the headline percentage.
Where PayEurasia fits
PayEurasia operates local collection and payout rails across Bangladesh, India, Pakistan and Nepal behind one API, one reconciliation model and one settlement relationship, with provider redundancy so a single acquirer incident degrades performance instead of stopping payments. The high-risk payment solution for Nepal page describes the local coverage, the API documentation covers authentication, webhooks and errors, and merchant onboarding lists what underwriting requires.
Related guides
Talk to PayEurasia
Working in a high-risk vertical across South Asia? We can probably help.
Request integration →