Tunisian exchange rates, with the receipt attached.
We read what Tunisian institutions publish: the central bank’s interbank, banknote and forward tables, and keep each one in the unit it was printed in. Alongside them sit precious metals and crypto as a separate, always-labelled class. Every value carries its source class, its capture time and the depth of history behind it.
A captured reading, not a live feed. Central-bank and global-market figures are real, read on a date we cannot confirm. Bank buy and sell sides are sample; no scraper runs yet, so the banks named have not quoted these numbers.
Beyond what a Tunisian bank will quote you.
Every currency the Banque Centrale de Tunisie quotes, in three tables: interbank, banknote and forward, each in its published unit.
Quoted per troy ounce as the market does, and per gram in 18k and 21k as Tunisia buys it.
Each charted only as far back as it actually goes, which ranges from 2022 to late 2025, and differs per asset.
Every number says what kind of number it is.
A Tunisian institution’s rate is a value someone printed, on a page we fetched, at a time we recorded. The date it was printed and the date we read it are both kept, because they are not the same day.
A global reference rate is a mid-market aggregate with no retail side and no institution page behind it. It carries its class, capture time and history depth, and is labelled everywhere it appears. We never blend the two.
Stop scraping banks yourself.
One app on your phone, or one endpoint, with a record of exactly where every figure came from. Read-only REST at /v1 and a domain MCP server at /mcp, both live. Keys and rate limits are specified and not yet built.