Why Machine Observer exists
Autonomous software increasingly discovers services, calls APIs, receives delegated authority, pays for resources, and interacts with markets. The evidence for that activity is fragmented across protocols, registries, vendor APIs, public datasets, independent probes and transaction systems. Machine Observer exists to turn that fragmented record into a source-traceable view of the machine economy without pretending unlike measurements are interchangeable.
What Machine Observer observes
Machine Observer maintains indexes and longitudinal measurements across machine commerce, agent-native web infrastructure, payments, identity and trust, APIs and data markets, compute, regulation, advanced manufacturing and other sectors where software is becoming an economic actor. The goal is to measure specific behaviors and infrastructure rather than reduce the entire machine economy to one headline number.
Evidence first, interpretation second
The operating model is observation → provenance → claim → derived signal → index → research conclusion. Source links stay attached to claims, vendor declarations remain distinct from direct observations, and historical gaps remain gaps rather than being silently interpolated. Research can change when the evidence changes while the underlying timestamped record remains inspectable.
How humans use it
For human readers, the public site provides indexes, research, history, source health, methodology and sector views. The useful path is usually to start with an index, inspect the evidence behind a surprising signal, then read the research explaining what the measurement may mean and where the interpretation stops.
How agents use it
For software, Machine Observer exposes machine-readable JSON, discovery manifests, OpenAPI metadata, task routing and MCP-compatible surfaces. Free discovery and evidence checks come first. Paid x402 products are quoted when a task requires deeper analysis or a specialized dataset, but discovery never authorizes spending: the buyer or its principal must independently approve any payment under its own wallet policy.
What Machine Observer is not
Machine Observer is not a claim that every autonomous system can be observed, and no individual index is presented as a complete market size, safety score or universal measure of agent capability. Commercial products, sponsorship and advertising are separated from editorial conclusions and source confidence. The value of the observatory depends on making uncertainty, provenance and coverage boundaries visible.
What the measured record says
Agent Economy tracked 176,520,376 x402 transactions and $41,547,691 in cumulative settled value as of its August 31 refresh.
QuickNode’s ERC-8004 Explorer showed 116,377 agent registrations and 29,831 feedback events over the trailing 30 days when checked September 1.
APIs.guru reported 2,529 APIs, 3,992 specifications and 108,837 endpoints in its public directory metrics.
The Agentic Payments Landscape cataloged 33 entries across substrate, identity and trust, authorization, commerce and checkout, and settlement and rails.
Use the live data
This guide explains the concepts; the underlying index continues to change. Open the related live index and methodology →
Interpretation boundary
Counts and probes are reported with their own populations and timestamps. Vendor declarations remain distinct from direct observations; no individual measurement is presented as a total market size or universal trust verdict.
Read the Machine Observer methodology →
For machine readers
sector: machine-commerce
last_updated: 2026-10-07
claim_ids: MO-2026-0001, MO-2026-0004, MO-2026-0005, MO-2026-0008
structured_representation: /api/v1/stories?slug=what-is-machine-observer
claim_endpoint: /api/v1/claims?id=<CLAIM_ID>
historical_record: /api/v1/history/observations