Data

ANP — Brazilian fuel-price weekly surveys

Direct access to ANP's weekly fuel-price surveys (gasoline, ethanol, diesel) at national, state, capital, and municipal levels. Returns typed responses with embedded data_source_url + survey_week + last_collected_at so agents can self-attribute provenance inline. Covers ~5000+ Brazilian municipalities with weekly cadence.

Verified: 2026-05-03 (laranja-data-plane-audit-2026-05-02)

Risk Flags

  • LOW operational ANP's weekly XLSX shape can change between editions; the parser is lenient but new columns may not be exposed until the wrapper updates.
  • LOW regulatory All ANP data is public open data. Safe for high-volume integration.

Cost

Type: Free · Free tier: Unlimited via auxiliar.ai gateway. ANP upstream is open data.

Hidden costs

  • ANP weekly XLSX is fetched + parsed; first call per week takes ~10-15s; cached after.
  • Survey cadence is weekly (publishes Tuesdays); freshness is bounded to ~7 days.

Install

Default

curl -s -H "Authorization: Bearer $AUXILIAR_TOKEN" "https://api.auxiliar.ai/api/invoke/anp_combustiveis_precos_municipio?summary_url=...&uf=SP&municipio=SAO+PAULO"
# Latest survey week (no params):
#   GET /api/invoke/anp_combustiveis_ultima_semana
#     → returns the most recent week metadata + summary_url for the XLSX
#
# Prices by municipality:
#   GET /api/invoke/anp_combustiveis_precos_municipio
#     params: summary_url (from ultima_semana), uf (2 letters), municipio (name)
#     returns: { uf, municipio, total, rows, data_source_url, survey_week, last_collected_at }
#
# National-level time series:
#   GET /api/invoke/anp_combustiveis_media_nacional
#     params: produto (default "GASOLINA COMUM"), last_weeks (default 8)
#     returns: weekly average time series with same self-attribution fields
#
# All responses include `data_source_url` (the original ANP XLSX URL),
# `data_source_authority: "anp.gov.br"`, `survey_week`, and
# `last_collected_at` (UTC ISO timestamp) — quote these inline in
# the agent's answer for clean source attribution.

Estimated time to first success: ~5 min

Dependencies

Minimum runtime: Any HTTP client

Composes with: Banco Central do Brasil — monetary policy + indicators

Distribution

Repository
https://github.com/Tlalvarez/Auxiliar-ai