Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.simplefunctions.dev/llms.txt

Use this file to discover all available pages before exploring further.

SF persists most of its compute output to long-term storage so research and audit are reproducible.

​
R2 archives

Daily themed dumps to Cloudflare R2:
  • data-dump-markets — market_price_snapshots, market_indicators, market_regimes
  • data-dump-compute — every compute table (33 of 37 tables)
  • data-dump-thesis — thesis_videos, cross_venue_pairs, legislation_market_pairs
  • data-dump-obs — observability tables (cron_run_log, health_alerts, cost_events)
Crons under src/app/api/cron/data-dump-*. Bucket: sf-data-dump.

​
HuggingFace datasets

Public datasets at huggingface.co/datasets/SimpleFunctions/*:
  • sf-index-history — full SF Index history.
  • sf-settled-markets — resolved markets (volume > $10K filter).
  • sf-calibration-scorecards — model calibration over time.
  • sf-world-state-history — daily world snapshots.
Crons: hf-export, hf-settled-markets, hf-sf-index-history, hf-calibration, hf-backfill-world-state. They write via lib/hf-client.ts which supports LFS for files >10 MiB.

​
Wayback / Common Crawl

wayback-submit-core and wayback-submit-answer push canonical SF pages to Wayback Machine for citation tracking. citation-monitor polls back to see who’s citing us.

​
Provenance

Every dump row carries the trace_id from the producing cron. The dump_at timestamp is embedded in the file path.

​
Access

# R2
aws s3 ls s3://sf-data-dump/markets/2026-04-30/

# HF
curl https://huggingface.co/datasets/SimpleFunctions/sf-index-history/resolve/main/index.parquet
R2 access requires SF API key + a separate sf data-dump grant flow (TBD).

​
Next steps

Provenance

trace_id in every dump row.

Surface map

Full table list and surface inventory.