Centralized RPC providers can become choke points that silently filter or delay transactions, so detecting censorship requires both network-aware measurement and cryptographic confirmation. Mempool divergence analysis compares pending transaction visibility across multiple endpoints to reveal omissions. Submitting identical low-fee, non-conflicting probe transactions to several RPCs and observing which mempools list them exposes selective filtering, while tracking the time from submission to block inclusion quantifies suppression. Vitalik Buterin at the Ethereum Foundation has written about how transaction visibility and miner incentives affect inclusion, framing why such probes are relevant.
Network- and application-level probes
Observing gossip-layer propagation complements RPC checks because a transaction visible in peer-to-peer propagation but absent from an RPC provider indicates server-side filtering rather than network loss. Passive packet capture from a personal node or participation in peer networks can show whether a transaction reached validators independent of the RPC. Gossip monitoring and independent peer sampling are effective for isolating where censorship occurs, though they require more infrastructure and technical expertise and may not be feasible for casual users.
Cryptographic receipts and relay comparison
Using private relays and MEV-aware channels provides a control against public RPC censorship. Philip Daian at Flashbots has developed private-relay mechanisms and research on MEV that illustrate how alternative submission paths bypass some forms of RPC filtering. Comparing outcomes from standard RPCs, Flashbots-style private submissions, and a self-hosted node gives strong evidence: if a transaction succeeds via a relay or local node but fails through a commercial RPC, censorship is likely. Collecting signed inclusion receipts and transaction proofs from validators creates immutable evidence for auditing.
Causal factors include commercial incentives to comply with regulators, risk-avoidance policies, or contractual relationships with validators that encourage content-based filtering. Consequences reach beyond technical performance: censorship undermines trust in decentralized finance, concentrates power in providers that may reflect territorial regulations, and forces projects and users to adopt resource-intensive mitigations. Running a personal node is a robust anti-censorship strategy but carries hardware, maintenance, and energy costs that vary by region and community resources. Combining mempool divergence, gossip-layer observation, relay comparison, and cryptographic receipts produces the most reliable detection posture and yields auditable records to support accountability and policy responses.