Agoric RPC endpoints: free public URLs by latency
HTTP round-trip latency for the Tendermint `status` method against every free, no-key public Agoric Cosmos SDK endpoint, audited every 60 seconds from 3 regions.
TL;DR. As of , StakeAngle has the lowest median latency of the 3 free public Agoric RPC endpoints measured, 93 ms (p50, 24h, 3 regions). Source: OpenChainBench, https://openchainbench.com/benchmarks/agoric-rpc. 3 public endpoints, no key: agoric-rpc.stakeangle.com, agoric-rpc.polkachu.com, main.rpc.agoric.net. Full URLs with their 24h median are listed below.
Companion page
Comparing free RPC endpoints across every chain we measure? Open the cross-chain RPC matrix →
Agoric is a Cosmos SDK chain (agoric-3) running a JavaScript smart-contract environment. It is also this cluster's clearest illustration of why a list of endpoints is not a list of providers: the Cosmos chain registry publishes eleven public RPCs for Agoric and only a handful answered the Tendermint `status` probe on 2026-10-07. The eight that did not were kjnodes, AutoStake, Stakeflow, 0base.vc, w3coins, StakeWithUs, Stake&Relax and Bro_n_Bro. This page ranks the three that work: the Agoric official endpoint (main.rpc.agoric.net), StakeAngle and Polkachu.
Methodology
Agoric: 3 free public RPC endpoints answering Tendermint status every 60 seconds from us-east, eu-west and Singapore, ranked by 24h median. StakeAngle leads at 93 ms; the chain-official endpoint measures 213 ms. We measure the round-trip latency of a single Tendermint `status` call against every no-key public Agoric endpoint that sustains continuous probing, 3 providers measured, every 60 seconds, from us-east, eu-west and Singapore. The harness classifies every response (ok / http_err / jsonrpc_err / stale / timeout) with a Agoric-scaled staleness gap (20 blocks at Agoric's ~6 s block time). The cross-chain view lives on the parent `rpc-capabilities` benchmark; this page is the Agoric-scoped answer with per-region breakdowns as a first-class dimension.
Results: 3 free public Agoric RPC endpoints ranked by p50 latency (24h, 3 regions)
| № | Providers | p50 | p90 | p99 | Success |
|---|---|---|---|---|---|
| 1 | StakeAngle | 93 ms | 96 ms | 499 ms | 100.00% |
| 2 | Polkachu | 127 ms | 138 ms | 177 ms | 100.00% |
| 3 | Agoric | 213 ms | 433 ms | 762 ms | 100.00% |
Public Agoric RPC endpoints measured
The 3 no-key endpoints answering our probes today, with their current 24h median. Paste one into a wallet or a client as is: no signup, no key. Private providers (API key) are never listed with a URL.
StakeAnglehttps://agoric-rpc.stakeangle.com93 ms100.0%
Polkachuhttps://agoric-rpc.polkachu.com127 ms100.0%
Agorichttps://main.rpc.agoric.net213 ms100.0%
Median round-trip and success rate over the last 24 hours across the probe regions; the ranked table above carries p90, p99 and the per-region split.
Frequently asked
Which Agoric RPCs work without an API key?
3 endpoints are probed on this chain: Agoric (main.rpc.agoric.net), StakeAngle (agoric-rpc.stakeangle.com), Polkachu (agoric-rpc.polkachu.com). Every listed endpoint was live-verified with a Tendermint status call returning a parsable response before inclusion.
How is Agoric RPC latency measured here, technically?
One identical Tendermint status call every 60 seconds against each provider from each of 3 regions, using the same plain HTTP client the rest of the RPC cluster uses. Wall-clock round-trip is recorded at millisecond precision; p50/p90/p99 are computed via Prometheus `quantile_over_time` over 24 hours. Responses are classified (`ok` / `http_err` / `jsonrpc_err` / `stale` / `timeout`) so an endpoint stuck on an old head or returning errors behind HTTP 200 is never ranked as fastest.
The Cosmos registry lists eleven Agoric RPCs. Why are only three here?
Because eight of them did not answer. The registry is a contributed list and dead entries are rarely pruned, so copying it would publish endpoints that fail on first use. Every row on this page answered a real Tendermint `status` call before inclusion and is re-probed every 60 seconds from three regions; an endpoint that stops answering drops out of the ranking rather than sitting there with a stale number.
Source code github.com/ChainBench/OpenChainBench/tree/main/harnesses/rpc-capabilities