Evidence coverage = how much evidence stands under this classification — coverage, not prediction (Evidence Coverage v1, registered protocol). Score = composite research score. It is not an expected return, a probability, a price target, or a recommendation.

Labels are research classifications, not buy/sell recommendations:

STRONG_BULLISH · BULLISH · NEUTRAL · WATCH · WEAKENING · NEGATIVE_EVENT · BEARISH_WATCH · RISK_ALERT (each label links to its locked threshold definition)

There is no SELL or SHORT label. The SDK's _validate_label() is invoked on every signal-bearing return.

YUCLAW is the open evidence layer underneath AI research tools. Start with llms.txt and the machine-readable evidence_index.json (every page, packet, and protocol with stable URLs and data-through dates). Packets carry derived statistics, event CSVs, engine run JSONs, and citation snippets; yuclaw replay-lab re-computes the published statistics from the public bundle. Derived data only — preserve the disclaimers when quoting; nothing here is advice or a recommendation.

Version 9 adds sentence inspection, brief composition, tracked edits and translations, imported provenance records, portable verification and operation counts. Start with the English or French guide. It does not include live model-generation or watermark-detection connectors.

This website is the research-content site: evidence, labs and records you can read and re-compute. The v8 workbench is a different thing: an application that runs on your own computer (it binds 127.0.0.1 only; your data stays in the folder you name; there is no hosted service and no account). It traces one financial commitment from an exact source passage through a typed, frozen claim, its revisions, exact calculation, history and review to an export that a fresh workspace re-verifies. Four modules work on the same records: SHD Distillation Shield (protected evidence intake), EVO Evolution Evidence Audit (what changed in an AI system, and which evidence still applies), COM Research Commons Guard (a fair, bounded review queue) and PRC Independent Practice (attempt first, then compare).

pip install yuclaw
yuclaw workbench selftest                                  # checks this installation in a temporary fictional workspace
yuclaw workbench principals init --workspace ~/yuclaw-ws   # optional: roles for the four modules; prints the first administrator's credential once
yuclaw workbench serve --workspace ~/yuclaw-ws             # then open http://127.0.0.1:8765  (Ctrl-C stops it)
yuclaw workbench guide                                     # the full guide, from the package

In the browser: load a fictional example on the Workspace page, follow its seven steps, build the export, then verify it in a second, fresh workspace (current guide). Status: experimental, local, owner-operated. Including the modules activates nothing — no role, approval, budget or study exists until you set one up. Protected SHD admission needs Linux with Landlock; on macOS, Windows or a kernel without it that one route stays closed and everything else works. No independent security review has been performed, and no user study: human benefit is PENDING. What has and has not been demonstrated is on the evidence scoreboard, under its own as-of time.

pip install yuclaw
yuclaw demo                         # 3-minute guided "Why AMD?" journey
yuclaw why AMD --as-of 2026-05-20   # bundled offline signal
yuclaw verify AMD --date 2026-05-20 # check the ledger record
# all tickers/dates: connect the local backend — see README

SDK + REST API + MCP server documented at github.com/YuClawLab/yuclaw-brain. REST API terms at /API_TERMS.md.