GetBlock RPC competitor monitor — 2026-10-27
Criterion of the day: AI search visibility and LLM citations: documentation/indexing that gets cited by ChatGPT, Perplexity, Gemini and AI agents (ai-search-llm-citations).
Executive summary
Today’s pass separates AI-citation readiness from actual named-engine citation proof. Direct vendor evidence confirms multiple machine-readable documentation and agent-entry surfaces, but no monitored source demonstrates measured citation frequency in ChatGPT, Perplexity, or Gemini. The cumulative topic brief was therefore rebuilt around five fixed dimensions and all 15 tracked providers.
The strongest decision-relevant evidence is Alchemy’s explicit site/docs llms.txt, llms-full.txt, sitemap, OpenAPI, Markdown twins, and agent tooling. Infura, Chainstack, Ankr, Moralis, Blockdaemon, and Tatum also expose directly documented LLM-oriented surfaces. Changes in page hashes or navigation alone were not treated as market changes.
Confirmed findings and matrix changes
- Alchemy — expanded to 4/5 dimensions. A first-party 404 response explicitly points agents to site llms.txt, llms-full.txt, sitemap.xml, openapi.json, docs
llms.txt, and Markdown twins; Alchemy docs also advertises agent skills, MCP, and CLI tooling. This is direct discovery/readiness evidence, not citation-share proof. - Infura, Chainstack, Ankr, Moralis, Blockdaemon, and Tatum — normalized from direct docs evidence. Confirmed combinations include
llms.txt, documented Markdown delivery,llms-full.txt, Ask AI/Open in ChatGPT, or MCP/agent navigation. Unsupported carry-forward claims were removed rather than preserved. - NOWNodes, dRPC, Pocket Network, and Lava — agent entry points only. NOWNodes publicly lists MCP Servers in pricing; dRPC advertises AI-agent skills/MCP on its home page; Pocket Network targets paid agent services and a Claude build workflow; Lava mentions AI-agent traffic. None of these pages proves named-engine citations.
Attribution and lifecycle corrections
- Blast API: the first-party site says the service is deprecated, and the docs resolve to Alchemy. The topic matrix now marks Blast dimensions
Not applicable — service deprecated; Alchemy documentation features are not attributed to Blast. - Pocket Network / Grove: Grove says it is winding down, while Pocket Network continues separately with an agent-service proposition. The matrix attributes current agent evidence specifically to Pocket Network.
- QuickNode: current changed pages contain substantial platform/navigation copy, but the captured evidence does not explicitly establish
llms.txt, Markdown delivery, structured AI indexes, or named-engine citations. Previous weakly sourced positives were replaced withNot verified in monitored public evidence.
Noise versus decision-relevant change
changed_vs_last=truefor QuickNode, Alchemy, Moralis, NOWNodes home, dRPC chainlist, Ankr docs, and the Infura blog redirect is a collection signal only. No claim was updated from a hash change alone.- QuickNode’s changed home/pricing/enterprise/blog/chain pages largely concern product, pricing, chain coverage, and navigation; they are off-criterion unless they explicitly document AI-discovery surfaces.
- Moralis’ broad site refresh is mostly product/navigation repetition. Only the direct docs wording for
llms.txt, the assistant, and “Open in ChatGPT” was promoted into the matrix. - Infura’s old blog URL redirects to MetaMask developer news; that redirect is not evidence of Infura AI-search visibility.
Provider-by-provider evidence ledger
| Provider | Decision-relevant evidence on today’s criterion | Limitation / status |
|---|---|---|
| QuickNode | No explicit criterion-specific feature confirmed in monitored pages | Changelog 404; changed pages treated as noise unless directly relevant |
| Alchemy | llms.txt, llms-full.txt, sitemap, OpenAPI, Markdown twins, agent skills/MCP/CLI | No named-engine citation benchmark |
| Infura | Docs advertise llms.txt, .md representation, and “Ask AI” | Blog redirects to MetaMask; no structured corpus/OpenAPI confirmed |
| Chainstack | Docs advertise llms.txt, llms-full.txt, and MCP navigation | Markdown representation not explicitly confirmed in capture |
| Ankr | Docs advertise llms.txt, Markdown delivery, Agentic RPC/MCP/TORPC | Structured full corpus/OpenAPI not confirmed |
| Moralis | Docs advertise llms.txt, assistant, and “Open in ChatGPT” | Markdown/full-corpus/OpenAPI not confirmed |
| NOWNodes | Pricing lists MCP Servers | Monitored legacy Postman docs URL returned 404; no llms/Markdown evidence |
| dRPC | Home advertises AI-agent skills and MCP | No llms/Markdown/structured discovery evidence |
| Blast API | First-party deprecation notice; docs redirect to Alchemy | Deprecated; do not transfer Alchemy evidence to Blast |
| Blockdaemon | Docs advertise llms.txt and .md pages | Old RPC and pricing URLs returned 404; no AI entry point confirmed |
| Tatum | Docs advertise llms.txt, .md pages, and Ask AI | Old supported-blockchains URL returned 404; no structured corpus confirmed |
| Pocket Network / Grove | Pocket targets agent services/Claude workflow; Grove wind-down confirmed | Grove docs DNS failed; entities/lifecycle must be attributed separately |
| LlamaRPC | No retrievable evidence | Site returned 526; docs/status DNS failed—an extraction gap, not a product claim |
| Allnodes | No retrievable evidence | All five monitored URLs returned 403—an access gap, not evidence of absence |
| Lava Network | Home explicitly includes AI-agent traffic positioning | Two legacy paths returned 404; no llms/Markdown evidence confirmed |
Coverage and interpretation
- 11/15 providers have at least one directly evidenced AI-discovery/readiness dimension.
- 0/15 have complete evidence across all five dimensions because no monitored vendor source provides a reproducible ChatGPT/Perplexity/Gemini citation study.
- Alchemy has the broadest monitored publishing surface at 4/5. Infura, Chainstack, Ankr, and Tatum each have 3/5.
- Missing evidence is recorded as
Not verified in monitored public evidence, never inferred as “No.”
GetBlock implications
- Publish and maintain canonical site/docs
llms.txt, stable Markdown representations, a complete machine-readable docs corpus, sitemap, and OpenAPI surface. - Create explicit agent entry pages and task-oriented docs, but position them as retrieval readiness—not guaranteed LLM citations.
- Establish a recurring named-engine visibility benchmark with fixed prompts and recorded model/date/citation URLs. That would create differentiated proof competitors currently do not publicly show.
Publishing and automation status
- Pre-run encyclopedia publication reported success (
published_urls: 141, latest daily2026-10-27), but a post-edit rebuild is required and performed after this report/topic update. - Agentic routing paper update reported
ok=true,paper_updated=true, with 57 evidence items. The generated living paper URL is Agentic routing RPC. The routing-paper update concerns latency/SLA/routing evidence and is adjacent to, not proof of, today’s citation criterion.
Research priorities
- Run a controlled citation benchmark in ChatGPT, Perplexity, and Gemini rather than using publishing features as a proxy.
- Probe canonical AI-discovery paths directly for QuickNode, NOWNodes, dRPC, Pocket Network, and Lava.
- Resolve LlamaRPC and Allnodes extraction gaps through permissible first-party access paths.
- Verify structured full-corpus/OpenAPI availability for Ankr, Moralis, Blockdaemon, and Tatum.