MCP App Store
by KeenEthics
Bajaj Finance icon

Bajaj Finance MCP Server for Claude and ChatGPT

By Bajaj Finance Ltd

Overview

Connect Claude to your Bajaj Finance account to check your loans, dues, and EMI Network Card in plain language, without logging into the app or website. Sign in securely with your registered mobile number, then ask questions like "When is my next EMI due?" or "How much limit is left on my EMI Card?" With this connector, you can view your active loans with their status, EMI amounts, and outstanding dues; check Flexi loan details, including available withdrawal limits; see your EMI Network Card limit, status, and card details; and get answers about Bajaj Finance products, charges, and eligibility. The connector is read-only. It retrieves your information but never makes payments, changes your account, or submits applications on your behalf.

How to connect

  1. Claude: Open the Bajaj Finance listing in the Claude connectors directory and choose Connect. Claude listing Bajaj Finance exposes 11 tools here.
  2. ChatGPT: Open the Bajaj Finance listing in the ChatGPT app directory and choose Connect. ChatGPT listing Bajaj Finance exposes 11 tools here.

Tools

customer_details

ChatGPT
Profile of the customer — name, date of birth, gender, city, state, pin code, mobile and email. Answers questions like: What details do you have on file for me?

customer_relations

ChatGPT
The loans and EMI cards the customer holds with us, each with its status and key figures. The fastest single call to answer "what do I have?". Each loan carries its account number at loan[].agreementNo — that is the agreementNo every loan_* tool takes, so call this first when a question is about a specific loan. Empty lists mean the customer holds nothing of that kind.

emi_card_details

ChatGPT
The customer's EMI cards — limits, status, and the purchases made on each card. Answers questions like: How much limit is left on my EMI card? Is my EMI card active? What have I bought on my EMI card? Returns each card with its masked number, variant, name on the card, status, total credit limit, available limit, amount used, expiry date, whether documents are pending, and the block reason if the card is blocked — plus whether a repayment mandate is registered. Each card's loan_list holds the agreement numbers of the loans taken on that card, one per purchase. For anything about a purchase — its EMI, outstanding amount, dues or status — pass that agreement number as agreementNo to loan_summary or loan_dues. An empty or null loan_list means nothing has been bought on the card.

loan_dues

ChatGPT
Everything owed on one loan, everything already repaid, and what part payment is allowed. Answers questions like: Do I have any overdue amount? Do I have any penalties or bounce charges? How much have I paid so far? Returns what is owed (total dues, total overdue, EMI overdue and the date it fell due, principal and interest overdue), the charges (penalty due, bounce charges, other overdue charges), the total repaid so far, and part payment (minimum and maximum currently allowed). Pass agreementNo — the loan account number (LAN) — from customer_relations: loan[].agreementNo. A negative part-payment maximum means part payment is not currently permitted. AMC fields are null on products that do not carry a maintenance charge.

loan_flexi

ChatGPT
Flexi loan withdrawal limits and recent transactions. Answers questions like: How much can I withdraw? How much limit have I already used? What was my last withdrawal? Returns the limits (total withdrawable, amount already utilized, maximum currently available, available drawdown limit, maximum and minimum withdrawal amounts, and the reason for a drawdown block if one applies) and the recent transactions — each with its type (drawdown, part payment), amount, date, part-payment status and failure reason if any. Pass agreementNo — the loan account number (LAN) — from customer_relations: loan[].agreementNo. Applies to Flexi loans only: on any other loan every field here is zero or null. Check the Flexi flag via loan_summary if you are unsure.

loan_summary

ChatGPT
Overview of one loan — the loan itself and its EMI. Reach for this first when a customer asks about a specific loan. Answers questions like: What is my outstanding balance? What is my EMI amount? When is my next installment due? Returns the loan (agreement number, product, status, Flexi flag), the amounts (total loan amount, principal outstanding, interest rate), the EMI (monthly amount, next EMI amount and date), the tenure (remaining and total, in months) and the maturity date. Pass agreementNo — the loan account number (LAN) — from customer_relations: loan[].agreementNo. For anything owed or repaid use loan_dues. For Flexi withdrawal limits and transactions use loan_flexi.

product_constraints

ChatGPT
The rules that apply to one product — which charges are levied and what the product permits. Answers questions like: Are foreclosure charges applicable? Is part payment allowed? Are there maintenance charges? Returns a set of flags describing what applies to the product, so you can answer whether something is chargeable BEFORE quoting any figure. If you do not have a valid product_id, call product_master_list first. A null flag means the source did not state a rule — treat it as "check", never as "no". For what a specific customer owes on their own loan, use loan_dues.

product_faq

ChatGPT
Search the frequently asked questions and return the answers that fit what the customer asked. Answers questions like: What happens if I miss an EMI? Can I foreclose early? What documents do I need? QUERY — pass the customer's question roughly AS THEY ASKED IT. The index holds real customer questions, so natural phrasing matches best. Do NOT compress to keywords and do not pad it out. good "I want to foreclose my loan" good "I have paid my EMI but it is not updated in my statement" poor "foreclosure charges" keywords lose the context that disambiguates poor "Could you kindly tell me whether it is..." filler dilutes the match Trim the pleasantries, keep the sentence. READING RESULTS — score is a DISTANCE, so lower is closer. Under 0.30 is a strong match; 0.30-0.45 is usable; above 0.45 is drift — ignore it rather than answering from it. The top hit is normally the answer; do not blend several. PRODUCT — pass product_id whenever you know it. The same question is indexed once per product with a DIFFERENT answer each time, so an unnarrowed search can hand you another product's rule. When you know the product, product_info usually answers better than this — it reads that product's own document. Use this one for general questions, or when no product was named. For rules on charges and what is permitted, use product_constraints.

product_info

ChatGPT
Search one product's knowledge document and return the passages that answer the question. PREFER THIS over product_faq once you know which product the customer means — it reads that product's own document and answers in more depth. Answers questions like: What is the interest rate on this product? What is the maximum tenure? What is covered? QUERY — natural phrasing, as the customer asked. Each block in the document is indexed together with the ways customers ask for it, so a real question matches best. good "can I get part of my pledged gold released" good "are there foreclosure charges" poor "partial release pledged gold" keywords can land on the wrong block entirely Trim the pleasantries, keep the sentence. READING RESULTS — score is a DISTANCE, lower is closer. Under 0.45 is a good match here; above 0.55 is drift. Take the top hit. product_id is REQUIRED. If you do not have a valid one, call product_master_list first — a guessed id returns nothing, which reads as "we do not have that" when we do. For questions spanning products, or when no product was named, use product_faq.

product_master_list

ChatGPT
Look up the valid values for a product input you do not already know, from Bajaj Finance's own product catalogue. Other tools name their lookup path in the input description — for example "[lookup: product.id]". Pass that path here to get the values that input accepts. path picks one of a fixed set of lists — it is not an API endpoint, and no request is built from it. Add a query to narrow the list: the names are searched, so the customer's own wording usually finds the right entry ("gold" finds Gold loan). ALWAYS call this before product_info or product_constraints unless you are certain of the exact id. Ids are snake_case and not guessable — personal_flexi_loan, bajaj_insta_emi_card, loan_against_fixed_deposit. A guessed id returns nothing rather than an error, so it reads as "we do not have that" when we do. customer says "gold loan" -> query "gold" -> gold_loan customer says "EMI card" -> query "emi" -> bajaj_insta_emi_card Paths: product.id | product.type

show_emi_card

ChatGPT
Shows the customer's EMI card as a card visual — masked number, name, status, expiry and available limit. Use it when the customer asks to see their EMI card. To answer a question about the card's figures, use emi_card_details instead.

customer_details

Claude

customer_relations

Claude

emi_card_details

Claude

loan_dues

Claude

loan_flexi

Claude

loan_summary

Claude

product_constraints

Claude

product_faq

Claude

product_info

Claude

product_master_list

Claude

show_emi_card

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/bajaj-finance"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/bajaj-finance)