CTO
Live in beta2025 — presentChimère
Instant cross-border payments for Indian freelancers. A US client pays in USD, rupees land in the freelancer's bank account in 20 seconds, at 1% all-in with no hidden forex markup.
Visit ChimèreThe corridor is the market, the automated compliance is the difficult part, and the MCP endpoint is why the payer doesn't have to be a person at all.
- Role
- CTO
- Scope
- Settlement rails, compliance engine, agent layer
- Founder & CEO
- Tanishq Wadhwani
- Corridor
- United States → India
- Headline claim
- 20 seconds, 1% all-in
The problem
Money that arrives lighter than it left
An Indian freelancer invoices a US client. The client pays on time, in dollars, from an ordinary US bank account. Then nothing happens for a while.
The dollars do not travel to India. They move between banks that hold accounts with one another, and where no such pair exists the payment is handed along a chain of intermediaries until one is found. Each bank in that chain takes a fee, and each takes its own time. Somewhere along the way the dollars become rupees at a rate the freelancer never sees quoted, because the margin is priced into the rate rather than charged as a line item. The invoice said one number. The bank statement says a smaller one, and nobody sends an explanation.
The delay costs more than the fee does. Money in transit cannot pay rent or fund next month's work. A freelancer running on invoices is running on a cash cycle, and several days of float in the middle of it decide whether the next project can be taken on at all.
Time from the client paying to rupees in the account
- Wire through correspondent banks
- Days
- Chimère
- 20 seconds
What it does
Three steps for the freelancer, twenty seconds for the money
- 01
Create and send a payment link
The freelancer raises a link against the invoice and sends it to the client.
Freelancer
- 02
The client pays in USD
Payment is made in dollars. That is the whole of the client's involvement.
USD in
- 03
INR lands in the bank account
Funds settle in real time the moment they arrive, and rupees reach the freelancer's own account.
20 seconds
- Settlement
20 sec
Settlement
Funds settle in real time the moment they arrive.
- Total cost
1%
Total cost
All-in, with zero hidden forex markups.
- Held natively
USD
Held natively
Dollars can sit in an EEFC account rather than converting on arrival.
- Per-invoice limits
No caps
Per-invoice limits
Any amount, with no per-invoice ceiling.
Two of those matter more than they read. Capital stays liquid instead of sitting frozen while somebody's queue clears, and there is no per-invoice ceiling, so a large project does not have to be broken into smaller invoices to get through.
The real product
The hard part was never moving the money
India does not treat an inbound payment as a private matter between two parties. Every remittance arriving from abroad has to be classified — what it was for, under which category of service — and the receiving bank has to be able to evidence that classification. The classification is the purpose code. The evidence is the FIRA, the Foreign Inward Remittance Advice, which is the document a freelancer needs to show the money was export earnings and not something else.
Done the ordinary way, both are manual. A person decides the code, a bank issues the advice on its own schedule, and the payment sits behind that paperwork. This is where the days actually go. The transfer is not what takes a week; the file waiting in front of it is.
Chimère classifies the purpose code automatically and generates the FIRA instantly. That is the whole of the trick. Once the compliance step runs at the speed of the payment instead of the speed of a queue, twenty seconds stops being a stunt and becomes roughly how long the transaction should have taken all along.
Automating the paperwork is what compresses days into seconds. Moving the money was never the slow part.
AI-assisted purpose codes
Every inbound payment has to be classified for the regulator. Chimère classifies it automatically instead of asking a person to.
Instant FIRA
The Foreign Inward Remittance Advice, generated on the spot rather than requested from a bank and waited on.
EEFC integration
Native Exchange Earners' Foreign Currency accounts, so dollars can be held as dollars rather than converted the moment they land.
Agent-native
The payer doesn't have to be a person
A payment link assumes a person: someone who opens it, reads it and decides. The same rails also expose a native MCP endpoint, which puts the identical payment within reach of software that has no browser and no hands. An agent can be given the ability to pay in much the same way it is given the ability to search.
This is the part of Chimère that joins up with everything else I work on. I want agentic systems that are autonomous in the full sense — software with an identity of its own, able to hold something, earn, spend, and answer for what it did. Payment is where that stops being a design discussion. A payment either happened or it did not, it is attributable, and it is regulated. An agent that can pay has to be an entity the system is willing to name.
Money should move as fast as the thing that decided to spend it.
Chimère Pay MCP
A native endpoint that plugs the rails into any AI agent, rather than requiring the agent to impersonate a person at a checkout.
Chimère Pay Agent
An AI agent that handles a cross-border payment end to end, from raising the link to settled rupees.
Two classes of payer
An agent driving a browser borrows a human's interface and credentials. A native agent has neither and needs authority of its own. Those are two payment paths, not one path with a flag set.
How it fits together
One set of rails, two front doors
A human opening a payment link and an agent calling the MCP endpoint arrive at the same compliance step and the same settlement. Serving both is affordable precisely because only the entry point differs.
Entry points
Human · agent
A payment link the client opens in a browser, and an MCP endpoint an agent calls directly.
Capture
USD
The client pays in dollars, which can be held in a native EEFC account instead of being converted on arrival.
Compliance engine
The hard part
Purpose code classified automatically, FIRA generated instantly. This layer decides whether the money waits.
Settlement rails
20 seconds
Funds settle in real time the moment they arrive.
Payout
INR
Rupees reach the freelancer's own bank account, with no invoice cap.
What I think
Why this corridor, and what it is really about
The freelancer corridor is a wedge rather than the whole idea. US to India, freelancers first, because it is a real market with a specific complaint, and because it forces the compliance problem immediately instead of letting it be deferred to a later version. Once inbound India works, the harder half of the problem is already behind you.
I like problems where the answer has to be fast and correct, and where being approximately right is the same as being wrong. Payments are one of those. A classification that is usually right is a compliance failure. A settlement that is usually instant is not instant.
The direction from here is authority with limits — an agent given a budget and rules about where it may spend, with the classification and accounting that has to sit underneath. None of that is shipped, and I would not present it as though it were. What is live today is the corridor.