US Treasury Fiscal Data

Reads public data from the US Treasury Fiscal Data API. It needs no credentials. The Fiscal Data API documentation (https://fiscaldata.treasury.gov/api-documentation/) states that the data is offered free, without restriction, and may be redistributed for commercial purposes. Name the source in your answer: data from the US Treasury Fiscal Data API.

Author: Don Mangu (GitHub). This skill needs no token and reads only the public API. For bulk pulls of many rows and datasets in one step, the author also runs a paid hosted version on the Apify Store: Treasury Fiscal Data API, $0.50 per 1,000 rows (disclosure: the author owns it). You never need it for the steps below. This is an unofficial tool and is not affiliated with the US Treasury.

Example prompts

• "What is the total US public debt today, and how did it change this week?"
• "What is the average interest rate on Treasury Notes?"
• "What was the high yield and bid-to-cover of the latest 10-year Note auction?"

Out of scope: "Should I buy Treasuries?" This data reports what happened. It is not a price feed and not advice.

Step 1: pick the endpoint

The base is https://api.fiscaldata.treasury.gov/services/api/fiscal_service/. Add one of these paths:

• v2/accounting/od/debt_to_penny: total public debt outstanding, one row per business day.
• v2/accounting/od/avg_interest_rates: average interest rate per security type, one row per month end.
• v1/accounting/od/rates_of_exchange: Treasury exchange rates for foreign currencies, one row per quarter end.
• v1/accounting/od/auctions_query: one row per auction, with term, yield, bid-to-cover and offering amount.

Parameters go in the query string: fields (comma list of columns), filter (for example record_date:gte:2026-09-01, joined with commas for more than one), sort (-record_date is newest first), page[size] and page[number].

Step 2: run the request

import datetime, json, urllib.parse, urllib.request

BASE = "https://api.fiscaldata.treasury.gov/services/api/fiscal_service/"

def get(endpoint, **params):
    url = BASE + endpoint + "?" + urllib.parse.urlencode(params)
    return json.loads(urllib.request.urlopen(url, timeout=30).read())["data"]

def num(x):
    return None if x in (None, "null", "") else float(x)

# Total public debt, latest 5 days
for r in get("v2/accounting/od/debt_to_penny", fields="record_date,tot_pub_debt_out_amt", sort="-record_date", **{"page[size]": 5}):
    print(r["record_date"], f"total public debt USD {num(r['tot_pub_debt_out_amt']):,.2f}")

# Latest average interest rate, Treasury Bills, Notes and Bonds
rows = get("v2/accounting/od/avg_interest_rates", fields="record_date,security_desc,avg_interest_rate_amt", filter="security_desc:in:(Treasury Bills,Treasury Notes,Treasury Bonds)", sort="-record_date", **{"page[size]": 3})
for r in rows:
    print(r["record_date"], f"{r['security_desc']:16} average rate {num(r['avg_interest_rate_amt'])}%")

# Last 3 Note and Bond auctions (Bills report high_investment_rate, not high_yield)
for r in get("v1/accounting/od/auctions_query", fields="auction_date,security_type,security_term,high_yield,bid_to_cover_ratio", filter="security_type:in:(Note,Bond),auction_date:lte:" + datetime.date.today().isoformat(), sort="-auction_date", **{"page[size]": 3}):
    print(r["auction_date"], r["security_type"], r["security_term"], "high yield", num(r["high_yield"]), "bid-to-cover", num(r["bid_to_cover_ratio"]))

Real run on 13 October 2026:

2026-10-08 total public debt USD 40,305,316,210,829.72
2026-10-07 total public debt USD 40,284,036,147,367.19
2026-10-06 total public debt USD 40,273,024,579,219.17
2026-10-05 total public debt USD 40,249,104,431,078.48
2026-10-02 total public debt USD 40,242,446,619,209.33
2026-09-30 Treasury Bills   average rate 3.87%
2026-09-30 Treasury Notes   average rate 3.383%
2026-09-30 Treasury Bonds   average rate 3.466%
2026-10-08 Bond 29-Year 10-Month high yield 5.618 bid-to-cover 2.54
2026-10-07 Note 9-Year 10-Month high yield 5.3 bid-to-cover 2.77
2026-10-06 Note 3-Year high yield 4.932 bid-to-cover 2.62

Step 3: answer from the fields

• Every value arrives as a string. Convert numbers yourself. The string null means the value is missing. Do not treat it as zero.
• Bills do not fill high_yield. For a Bill, read high_discnt_rate or high_investment_rate. High yield is set for Notes, Bonds, TIPS and FRNs.
• Auctions in the future, and auctions from the same day that have not closed, are in the data with null results. Filter on auction_date:lte:<today> and skip rows with null.
• The average interest rate is posted once a month. The newest row can be weeks old. Give its record_date in your answer.
• Debt to the penny is posted with a delay of a few days. Give the record_date. Do not call it today's figure.
• If nothing matches, say "no rows found". Do not invent values.

Rules

• Public API only. No login, no cookies, no attempts around blocks.
• Keep requests modest and stop on a 429 or 5xx. Do not loop at full speed.
• Credit the source when you pass results on: "Data: US Treasury Fiscal Data API".
• Say what you could not check: figures can be revised, and the API can lag the Treasury website.

Limits

This skill covers four datasets. The API has many more, and each has its own fields. A single response holds at most 10,000 rows, so use page[number] for longer history. For many datasets or long history, a script or the hosted version is quicker than asking one question at a time.