Launch issue. 10 vendors ranked across 4 form factors. Manus tops the inaugural composite; n8n lands a -41% MoM traffic warning that frames the whole issue.

What this index measures

The Fieldtested Power Index ranks 10 AI agent platforms on a 0–10 composite drawn from four reproducible sub-metrics — Authority Score, Organic Traffic, Backlinks (referring domains), and Growth (MoM organic traffic delta). All inputs come from Semrush US-database snapshots captured 2026-05-29. The fifth sub-metric in the published methodology — Search Interest via Google Trends — is reserved for v1.1; this launch issue redistributes that 0.20 weight evenly across the four available metrics (0.25 each).

The table above ranks each vendor by composite. Below: what the numbers actually say.

The headline: Manus arrives at #1

Manus (manus.im) takes the inaugural top spot at 6.27. For an agent platform that was barely on the Western radar 12 months ago, this is the most consequential data point in the issue. The composite is driven by a near-maximum backlinks sub-score (99.5/100 — the most referring domains in the basket after n8n) and a near-maximum traffic sub-score (98.3/100). The MoM growth is slightly negative (-5.2%) but off a base most Western competitors haven’t reached.

The Manus story is operationally important for B2B buyers in two ways. First, the platform’s autonomous task execution model is closer to what most procurement decks describe when they say “AI agent” than what most Western tools actually deliver. Second, the data residency and regulatory posture for Chinese-built agent platforms is unsettled in most enterprise compliance frameworks — meaning Manus’s market presence isn’t directly accessible to most regulated B2B operators today. The composite captures momentum, not deployability.

The framework layer is consolidating

LangChain at #2 (6.17) and Crew AI at #6 (5.25) bracket the framework category. LangChain still owns the long-tail SEO real estate that comes from being the de-facto agent tutorial reference for two years — 173k organic traffic, 24k referring domains. Crew’s composite is a third lower despite a more elegant multi-agent abstraction, because Crew’s discovery still happens through GitHub stars and developer word of mouth, not through general search.

The implication for B2B operators: LangChain remains the right default when “the team can read Python,” not because the framework is technically superior, but because the documentation surface and Stack Overflow depth are larger. Crew earns its slot when the work decomposes cleanly into roles. (Review: Crew AI.)

iPaaS: n8n drops -41% MoM and still ranks #3

n8n’s June composite is 5.76 — third place — but the underlying signal is a single outlier: organic traffic dropped from 393k to 231k between April and May. The composite math weights the absolute (231k traffic = max sub-score of 100) more than the delta (growth sub-score of 5.84/100, the lowest in the basket). On a longer time frame n8n is still the iPaaS leader; on a one-month window this is the largest single move in the issue.

Three plausible explanations: a Google algorithm impact on a specific content cluster (their integration directory drives most non-brand traffic), a competitive event we’re not seeing yet, or a measurement artifact at the Semrush layer. The next issue will tell us which. For now: don’t reweight your iPaaS evaluation based on one month of data, but watch this number. (Review: n8n.)

Managed SaaS: Lindy clearly leads, Relevance trails

Lindy (#4, 5.63) and Relevance AI (#9, 4.90) compete in the same Managed SaaS form factor and the gap is 0.73 composite points — wider than any other intra-category gap. The driver: Lindy’s organic traffic is 6× Relevance’s, and Lindy’s referring-domain count is 42% higher. The two products target overlapping audiences (sales/RevOps B2B teams) but Lindy’s broader use-case template gallery — captured in the review — has translated into broader organic search capture.

The directional read: in form factors where a single product can win the “fastest first-deploy” race against non-technical buyers, that single product captures most of the search-discovery upside. Relevance still wins specific revenue-team use cases, but the composite gap is structural, not cyclical.

Voice agents: tightly clustered, no clear winner

Vapi (#7, 5.05), Retell (#8, 5.02), and Bland (#10, 4.90) sit within a 0.15-point spread. All three operate the same form factor (voice-only, sub-second latency, programmable call flows) at comparable scale. None has emerged with structural advantages on any of the four sub-metrics tracked here. Bland’s -9.3% MoM growth is the only differentiator in this cluster, and it’s not large enough to move ranking.

Practical takeaway: for voice agent procurement in mid-2026, the decision should still come from latency tests, telephony provider compatibility, and compliance fit — not from this index. The Power Index doesn’t yet distinguish meaningfully between the three. (Review: Retell AI.)

Marketplace: Agent.ai grows +44%, ranks #5

Agent.ai’s growth sub-score (62.7/100) is the highest in the basket — and it’s the only sub-score where Agent.ai materially competes, since absolute traffic and backlinks remain modest. The +44% MoM growth reflects the network effect that Dharmesh Shah’s profile-driven community model is supposed to produce; it’s now visible in the search data, not just the founder’s posts.

The composite of 5.37 ranks #5 in the issue. Whether that holds in subsequent months depends entirely on whether the community converts curiosity into return visits. Marketplaces decay fast when discovery slows. (Review: Agent.ai.)

What’s next

Issue #2 (July 2026) will introduce Google Trends-derived Search Interest as the fifth sub-metric and update the methodology version to v1.1. Expect mild composite shifts as Trends data adds signal that current four-metric scores miss — particularly for brand-led products (Lindy, Relevance) versus organic-traffic-heavy ones (n8n, LangChain).

Vendor universe will expand from 10 to ~25 once Issue #2 includes form-factor-segmented sub-rankings (separate composites per form factor), starting with the iPaaS and Managed SaaS categories that have enough comparable entrants.

Reproducibility

Raw data: /data/indices/agents/2026-06.json. Methodology v1.0: /agents/methodology/. Re-computation from Semrush API + the published weights produces the exact composite scores in this table. Anomalies — particularly n8n’s MoM traffic delta — invite independent verification.

#	Vendor	Composite	Authority	Traffic	Backlinks	Trends	Growth	Δ
1	Manus	6.27	23	187,605	28,254	—	-5.2%	new
2	LangChain	6.17	22	173,249	24,487	—	-7.3%	new
3	n8n	5.76	25	231,134	29,796	—	-41.2%	new
4	Lindy	5.63	17	81,045	7,298	—	-3.4%	new
5	Agent.ai	5.37	0	8,887	3,293	—	+44.1%	new
6	Crew AI	5.25	8	24,636	7,022	—	+1.7%	new
7	Vapi	5.05	7	22,500	3,702	—	+1.2%	new
8	Retell AI	5.02	7	23,118	2,803	—	+3.2%	new
9	Relevance AI	4.90	3	13,449	5,124	—	-0.4%	new
10	Bland AI	4.90	7	22,686	4,077	—	-9.3%	new