Preflight & orchestration
Turns the question into an execution plan, estimates a conservative cost, chooses research depth, routes the right models/tools, and stops unnecessary spend.
AI4U TRUTH TASK FORCE · POWERED BY HARMONIA + DECISION LEDGER
Ask anything worth knowing.
Know the cost before you spend.
One evidence-first research system that helps you phrase the question, estimates the cost of answering it properly, decides how much research is justified, traces the sources, shows uncertainty, lets you challenge the result, and preserves every consequential change.
◆ Quote before research◆ Primary evidence first◆ Copies do not count twice◆ Challenges become separate cases◆ Decision Ledger preserves history
ONE SITE · FOUR PROOF-SEPARATED ENGINES
Simple for the user. Rigorous underneath.
The public sees one guided investigation. Underneath, each subsystem has one job and leaves evidence for the next.
Evidence assessment
Separates support from contradiction, tracks independent source roots, exposes uncertainty, and shows what can—and cannot—be reasonably concluded.
Adversarial correction
Any user can question a source, calculation, method, context, document or conclusion. The challenge becomes its own case and earns influence only through evidence.
Proof & history
Quotes, approvals, sources, research steps, score revisions, challenges, model/tool versions and final receipts remain reconstructable instead of silently changing.
THE PUBLIC WORKFLOW
From curiosity to defensible answer.
1AskType, paste, upload or speak what you want to know.
2ClarifyIf the question is vague, the site offers clearer alternatives without trapping the user.
3QuoteHarmonia predicts sources, searches, model passes, expected cost, safe quote and maximum spend.
4DecideLow-value work is screened out; worthwhile work is approved to a hard ceiling.
5ResearchPrimary records first, then independent support, contradiction, provenance and methodology.
6AssessThe case receives Evidence Support % ± Uncertainty—not a simplistic true/false label.
7ChallengeUsers can open separate challenge cases with or without bringing their own source.
8ReviseBetter evidence creates a new assessment revision; the old result remains inspectable.
9ProveDecision Ledger binds the quote, research, evidence, decisions, actual cost and outcome.
ECONOMIC DISCIPLINE
Research should earn the right to spend.
A major challenge to a primary dataset may deserve deep research. A side issue that could move the assessment by less than a point should not consume the same time or money. Harmonia estimates value before research and reconciles the quote against actual spend afterward.
POLITICAL-FIRST, UNIVERSAL
Politics is the stress test—not the boundary.
Political claims are adversarial, consequential and easy to weaponize, so the platform uses them to prove its neutrality and correction mechanisms. The same system can investigate science, business, media, consumer claims, historical questions, AI answers, photos, documents and more.
Claim, not personNo candidate “honesty score,” ideology ranking or voting recommendation.
Same standardEquivalent propositions receive equivalent evidentiary treatment regardless of speaker.
Serious-use gateNon-falsifiable, duplicate, off-topic or immaterial submissions do not consume deep-research resources.
Better evidence winsCorrections are expected. They create a new dated revision instead of erasing the old one.
TRY A COMPLETE CASE
Open the fictional unemployment investigation.
Inspect evidence roots, uncertainty, challenge handling, materiality triage and the revision model without making a real-world political judgment.
DESIGNED FOR ORDINARY PEOPLE
The user should not need to understand fact-checking to use the fact-checker.
Voice questionsGuided clarificationCost before spendEvidence receiptsSource lineagePhotos & videoChallengesSaved casesActual-cost reconciliationDecision history