Open Indian financial data as plain CSV, previewed before you use it.
Every value traces back to a source document kept unchanged in a public archive.
What SBI charges to convert between rupees and US dollars on a forex card, per the rate sheet it publishes each banking day. TT buy is what a dollar fetches; TT sell is what a dollar costs.
Rupees per US dollar on each business day since 1973, as the Bank for International Settlements publishes it. A reference rate rather than a price anyone transacts at; SBI's card rates are what a transfer costs.
India's consumer price index for all of India, rural and urban combined, as MoSPI publishes it on its 2024 base. Divide two months' values to get the inflation between them.
India's consumer price index back to 1957, which the IMF builds by linking India's earlier national series onto the 2024 = 100 base. It matches MoSPI's index from 2025; before 2013, it is the only monthly India CPI here.
Each series is one CSV with its full history, at a URL that does not change:
https://data.xfina.dev/v1/fx/bis-usd-inr.csv https://data.xfina.dev/v1/inflation/in-cpi.csv https://data.xfina.dev/v1/inflation/in-cpi-imf.csv
https://data.xfina.dev/v1/metadata.json lists every
series with its path, columns, source, licence, row count, date range, sha256 and last update;
read it to discover what is available and whether anything is new.
Downloads are compressed with Brotli or gzip when the client asks for it, which brings the larger series to about a quarter of their size (the BIS rates go from 254 KB to 65 KB). With curl:
curl --compressed -O https://data.xfina.dev/v1/fx/bis-usd-inr.csv
Over plain HTTP, ask with Accept-Encoding and the response says which one it used.
Browsers and most HTTP libraries do this on their own:
GET /v1/fx/bis-usd-inr.csv HTTP/2 Host: data.xfina.dev Accept-Encoding: br, gzip HTTP/2 200 content-type: text/csv; charset=utf-8 content-encoding: br access-control-allow-origin: *
Paths and columns under /v1/ are only ever added to, never renamed or removed.