ffiec-data-connect (Python)
Install
Installing ffiec-data-connect from PyPI. Python 3.11+ required.
Requirements
- Python 3.11 or newer. Required for pandas 3.0 and modern type hints.
- pip (or any PEP 517-compatible installer —
uv,poetry,pdmall work).
Windows users occasionally hit SSL certificate validation problems when talking to the FFIEC endpoint; if you see those, run in WSL, a Linux container, or Colab.
Install from PyPI
The current release is 3.0.0 — the first stable release on the 3.x line.
pip install ffiec-data-connect
# or pin the exact version
pip install ffiec-data-connect==3.0.0
This pulls ffiec-data-connect and its runtime dependencies (httpx, defusedxml,
pandas). polars is optional.
The 2.x line (last stable: 2.0.5) is still installable but is on a
different API — see Migration from v2
before pinning to it.
Optional extras
pip install 'ffiec-data-connect[polars]'
Adds polars for the output_type="polars" path. Polars is faster and has lower
memory overhead for large result sets; pandas remains the default.
Verify
import ffiec_data_connect
print(ffiec_data_connect.__version__)
If you see 3.0.0 (or newer), you’re set. If you see 2.x.x, you have an
older release — some of the calling conventions in these docs won’t apply;
see Migration from v2.
What got installed
The library is a thin typed wrapper over the FFIEC CDR REST API:
- Module:
ffiec_data_connect - Public API: re-exported from the top-level
__init__—OAuth2Credentials, thecollect_*functions,AsyncCompatibleClient,RateLimiter, and the exception hierarchy. See Library reference for the full list.
For AI coding assistants
Two machine-readable pointer files help LLMs find their way around:
call.report/llms.txt— an llms.txt manifest for the whole doc site, grouping every page by purpose so an AI agent can fetch just the relevant ones rather than crawling the full site.llms.txtin the library repo — a SOAP → REST refactoring guide with concrete find/replace rules for AI agents migrating v2 code. Shorter and more action-shaped than the full docs.
Next step
Authentication — get a JWT token from the FFIEC CDR portal.