fix: drop Moralis from indexing-freshness + data-api report Aug 2026

benchmarks/indexing-freshness.yml

@@ -2,9 +2,9 @@
 
 slug: indexing-freshness
 number: "070"
-title: Freshest wallet data API. Zerion, Moralis, Allium, GoldRush, Mobula
+title: Freshest wallet data API. Zerion, Allium, Mobula
 seo_title: "Freshest wallet data API 2026"
-seo_description: "How fast do wallet APIs index a new transaction? Zerion, Moralis, Allium, GoldRush, Mobula measured live on organic Base transfers, second by second."
+seo_description: "How fast do wallet APIs index a new transaction? Zerion, Allium and Mobula measured live on organic Base transfers, second by second."
 subtitle: Seconds between an organic transfer confirming on Base and the moment each wallet data API first returns it, measured continuously on real user transactions.
 category: Aggregators
 status: live
@@ -23,7 +23,7 @@ seo_intro: |
   transactions API until the tx hash appears. Because the ground
   truth is a random real transaction there is no benchmark wallet a
   provider could special-case, and every measurement is publicly
-  re-verifiable from the tx hash. Cohort. Zerion, Moralis, Allium,
+  re-verifiable from the tx hash. Cohort. Zerion, Allium,
   GoldRush (Covalent) and Mobula, all probed with the identical
   schedule from the same host. The timing context matters in 2026.
   Dune Sim is sunsetting and SimpleHash is gone, so teams are
@@ -59,7 +59,6 @@ methodology:
 findings:
   - "{{best_name}} currently leads at {{best_p50}} (p50 of found lags, 24h) across organic Base transfers."
   - "{{name:zerion}} sits at {{p50:zerion}}. Early runs showed a bimodal pattern, some events indexed in about a second and others surfacing several seconds later, consistent with a caching layer in front of the wallet endpoint."
-  - "{{name:moralis}} runs at {{p50:moralis}} with a notably tight distribution across events."
   - "{{name:goldrush}} trails at {{p50:goldrush}} on this probe. Its unified multi-chain schema trades freshness for breadth, a real trade-off teams should weigh explicitly."
   - "Miss rates matter more than medians: an API that fails to show a deposit within two minutes breaks the user flow entirely, so read the reliability column before the latency one."
 
@@ -104,17 +103,6 @@ providers:
       success: sum(increase(indexing_probe_total{provider="zerion", result="found"}[24h])) / sum(increase(indexing_probe_total{provider="zerion", result=~"found|missed"}[24h]))
       sample_size: sum(increase(indexing_probe_total{provider="zerion", result=~"found|missed"}[24h]))
       series: avg_over_time(indexing_freshness_seconds{provider="zerion"}[1h])
-  - slug: moralis
-    name: Moralis
-    tag: Web3 data API, wallet history across major EVM chains
-    formula: "Median (p50) over 24h of visibility lag in seconds between an organic Base transfer confirming and Moralis's wallet history endpoint first returning it, from the shared histogram estimator."
-    queries:
-      p50: histogram_quantile(0.5, sum by (le) (increase(indexing_freshness_seconds_histogram_bucket{provider="moralis"}[24h])))
-      p90: histogram_quantile(0.9, sum by (le) (increase(indexing_freshness_seconds_histogram_bucket{provider="moralis"}[24h])))
-      p99: histogram_quantile(0.99, sum by (le) (increase(indexing_freshness_seconds_histogram_bucket{provider="moralis"}[24h])))
-      success: sum(increase(indexing_probe_total{provider="moralis", result="found"}[24h])) / sum(increase(indexing_probe_total{provider="moralis", result=~"found|missed"}[24h]))
-      sample_size: sum(increase(indexing_probe_total{provider="moralis", result=~"found|missed"}[24h]))
-      series: avg_over_time(indexing_freshness_seconds{provider="moralis"}[1h])
   - slug: goldrush
     name: GoldRush
     tag: Covalent's multi-chain wallet API, unified schema

harnesses/indexing-freshness/cmd/script/config.go

@@ -23,12 +23,6 @@ type Provider struct {
 var providers = []Provider{
 	{Slug: "mobula", EveryN: 1, Budget: 250_000},
 	{Slug: "zerion", EveryN: 1, Budget: 55_000},
-	// Moralis free tier: 40k CU/day; /wallets/{addr}/history costs 25 CU
-	// per call, so full participation (144 events × 16 polls = 2,304
-	// calls = 57.6k CU/day) blows the daily cap ~2/3 through the day
-	// (the API then 401s until the UTC midnight reset). EveryN 2 keeps
-	// it at ~28.8k CU/day. Budget = 40k/25 CU × 30 days in calls.
-	{Slug: "moralis", EveryN: 2, Budget: 48_000},
 	{Slug: "goldrush", EveryN: 1, Budget: 90_000},
 	// Allium free tier: 20k calls/month, aggressive per-second limits.
 	{Slug: "allium", EveryN: 3, Budget: 18_000},

harnesses/indexing-freshness/cmd/script/pollers.go

@@ -46,12 +46,6 @@ func checkProvider(slug, wallet, txHash string) (bool, error) {
 			req.Header.Set("Authorization", "Basic "+base64.StdEncoding.EncodeToString([]byte(key+":")))
 			req.Header.Set("Accept", "application/json")
 		}
-	case "moralis":
-		u := fmt.Sprintf("https://deep-index.moralis.io/api/v2.2/wallets/%s/history?chain=%s&limit=20", wallet, chainSlug)
-		req, err = http.NewRequest("GET", u, nil)
-		if err == nil {
-			req.Header.Set("X-API-Key", key)
-		}
 	case "goldrush":
 		u := fmt.Sprintf("https://api.covalenthq.com/v1/%s-mainnet/address/%s/transactions_v3/?page-size=20", chainSlug, wallet)
 		req, err = http.NewRequest("GET", u, nil)