Find prior dead ends
Search by symbol, research generation, failure code or logic fingerprint before spending another compute cycle.
PUBLIC ALPHA · QFR 0.1
Good research also records what failed.
OpenQFR turns failed quantitative strategies into searchable, machine-readable evidence—so researchers and AI agents can avoid repeating the same dead ends.
NEGATIVE EVIDENCE ◆ REPRODUCIBLE RESEARCH ◆ PROVENANCE FIRST ◆ HUMAN + AGENT READABLE ◆ EXTERNALLY TESTED ◆ NO TRADING SIGNALS
01 / ARCHIVE
Search the evidence
Filter failed strategies by market or failure code. Every result links to its complete public record.
02 / WHY IT EXISTS
Failures are data, too.
A negative result can save the next researcher days of computation—when its assumptions, costs and provenance remain inspectable.
Inspect the evidence
Read the assumptions, validation scope, execution costs and outcome behind each failed hypothesis.
Contribute safely
Signed external records enter quarantine. Nothing becomes public until it has been reviewed.
03 / PUBLIC CHECKS
Externally tested. Publicly verifiable.
External community testers exercised the public A2A, MCP and REST interfaces. Findings, fixes and reruns remain open for anyone to inspect.
TRACEWAKE · 2026-09-09
Discovery inventory verified
The Agent Card and MCP surface the same three read-only capabilities. Public routes and tool discovery succeeded on rerun.
REPRODUCED AS FIXED ↗COMPACITAS · 2026-09-09A2A parity verified
A2A pagination and generation filtering were rerun against REST. Limits, totals and first-record ordering matched.
READ THE PUBLIC REPORT ↗Community interoperability checks are public test evidence—not certification, endorsement or a security audit.
04 / AGENT ACCESS
Built to be discovered.
Connect through A2A, MCP or REST. Public discovery is read-only and requires no credentials.