About Basis

The agent economy is growing fast, but its metrics are hard to compare. Agent counts, volumes, GDP figures, and skill catalogs are mostly self-reported, sybil-sensitive, or framed for maximum effect. Basis exists to separate claimed activity from verified activity.

We verify claims against onchain data, primary documents, and independent reporting; label them verified, partially verified, unverified, disputed, or stale; adjust populations for sybil risk with published heuristics; and anchor every report's hashes on Base so the record cannot quietly change.

Basis is adversarial but fair. We critique claims, not people. A disputed label means primary evidence conflicts with a statement — it is not an accusation of bad faith, and every project we cover has a standing invitation to correct the record with evidence.

No token. Basis will not launch one: an auditor whose own asset price depends on the audited ecosystem is structurally conflicted. Revenue comes from API access paid in USDC over x402, verification services, and diligence work — disclosed wherever it could matter. We hold no custody of user funds and take no positions in covered assets.

Roadmap. The claim database and audit reports are the foundation. Building on the same evidence substrate: Arena (continuous quality scores for live agent services), Clawscan (skill-risk scans for agent capabilities), Flightbox (runtime evidence and incident forensics), and Northstar (quality-ranked, facilitator-neutral discovery).