A public primary-source view of narrow observations at WagerX's MCP and A2A infrastructure. Repeated discovery polling is deduplicated from the headline activity count so automated scans do not masquerade as task-bearing requests.
The Observatory is a reusable public record of what reaches one WagerX service—not a market-size claim. It helps people building similar services and agents/aggregators cite a primary source, understand the measurement boundary, and compare later windows without turning private telemetry into a public log.
Use the selected window to see protocol mix, discovery versus task-bearing request classifications, registered capability demand, response performance and broad changes over time. These observations can inform capacity planning and instrumentation for a new agentic iGaming project.
The aggregate JSON is the machine-facing dataset for the same view. It includes the observation window, source and methodology links, the observation-through timestamp and the response generation timestamp.
All machine requests compared with task-bearing intelligence requests.
Successful production responses in the selected period.
Discovery traffic is shown separately from meaningful tool and A2A work.
Aggregated tool use only. Questions and caller identities remain private.
WagerX maintains a private historical dataset of privacy-protected interaction patterns, protocol activity, capability discovery, tool usage, response performance and historical trends. Selected aggregate research access may be considered for qualified researchers, agentic iGaming builders, AI companies, technology providers and data aggregators.
Methodology: Figures represent observed machine activity, not necessarily unique AI agents, companies or human users. Discovery methods (such as initialization and capability listing) are classified separately from task-bearing methods (such as tool calls and A2A messages); a task-bearing request is not proof that a human task completed successfully. Requests may include autonomous queries, capability discovery, repeated polling and protocol negotiation. The observation-through timestamp describes the latest source record included; the generated timestamp describes the JSON response, and neither is independent verification. Observations are made at WagerX and do not describe the whole ecosystem. Countries, raw prompts, IP addresses, ASNs, client identities and fine-grained individual activity records are not published.
