No motive claims
We do not infer fraud, abandonment or intent from silence. “Ghosted” is a discoverability label; formal classifications describe only verified public evidence.
About the project
Evidence before
epitaphs.
Dead Tokens is an investigative index of public project activity. It exists to replace rumor, stale lists and “inactive” shortcuts with inspectable sources, explicit uncertainty and reversible assessments.
Corrections are first-class
One verified newer event immediately reopens an assessment. Every correction should preserve the earlier snapshot and explain what changed.
Counterevidence stays visible
A project page is incomplete without the strongest fact against its current classification, alternative explanations and the next decisive check.
Precision is earned
Exact timestamps yield exact complete-day clocks. Dates yield ranges. Archives yield bounds. Missing histories remain unresolved.
Maintainable by design
A small evidence graph,
not a giant spreadsheet.
Provider rows map to one canonical project. Channels retain stable native IDs. Evidence events are immutable. Assessments are versioned outputs that can change without rewriting history.
PROJECT
↓identity · aliases · contracts · lineage
CHANNELS
ownership · status · coverage
PROVIDERS
IDs · market status · snapshot
EVIDENCE EVENTS
↓native ID · time interval · actor · hash · review
ASSESSMENT
clock · classification · confidence · unknowns
Appeal and correction standard
What changes a record?
- Identify the project.
Provide chain, contract or native asset identity—not only a ticker.
- Prove channel ownership.
Use a reciprocal official link, stable platform ID, signed notice or archived continuity.
- Link the native event.
A direct post, commit, release, governance item or dated first-party page is stronger than a screenshot or search snippet.
- Explain relevance.
Show why the event is meaningful project activity rather than automation, price promotion or an unaffiliated fork.
A future production release should expose a structured correction form, public revision history and reviewer attribution. Those features are deliberately not simulated in this static pilot.
Start with the evidence