What are the best alternatives to Relay.app?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Relay.app among agent builders are Taskade, Teneo, Base44, Gumloop. Each is graded on the same public evidence, and no vendor pays for placement.

Does Relay.app have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. Relay.app publishes public evidence of enterprise security and identity controls, covering a compliance posture such as SOC 2 or ISO 27001, single sign on (SSO or SAML), and role based access control (RBAC). Evidence basis: SOC 2 Type II and ISO/IEC 27001:2022 compliance, with a SOC 2 link published in the site footer alongside an information security page. NOTABLE GAP: no HIPAA compliance and no business associate agreement, identified by a reviewer as the sp. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Relay.app require human approval before an agent acts?

Full under the Agentic Index verification standard. Relay.app publishes public evidence of human approval steps and guardrails, so an agent action can be held for review before it executes. Evidence basis: The clearest case of human oversight as a design centre rather than an exception path in this index. Approvals, assigned tasks and human data entry can sit at any critical workflow stage, in flow approvals combine with branching logic so th. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can you bring your own model or route between providers with Relay.app?

Full under the Agentic Index verification standard. Relay.app publishes public evidence of bring your own model and bring your own key (BYOK), and of routing tasks between models or providers rather than being locked to one. Evidence basis: Native access to GPT-5.1 from OpenAI, Claude 4.5 Opus from Anthropic and Gemini 3 Pro from Google built into the platform WITHOUT requiring external API keys, so model choice sits with the builder and no separate provider accounts are neede. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Relay.app

Q: Who is Relay.app for?

operations and compliance teams, mid market SaaS, agencies and service firms, support, sales and finance teams

Also known as: Relay, relay.app, Relay.app Inc

Visit site

Agent builderindependentVerified 2026-08-05

Entry priceFree tier available; Team plan reported at $138 per month for 10 users; usage beyond included runs at about $0.05 per automation run; 50 percent annual discountFull pricing detail

AI agent and workflow platform built around human approval gates as a first class primitive, with built in GPT-5.1, Claude 4.5 Opus and Gemini 3 Pro, native data tables and all integrations included on every plan.

Relay.app is a workflow automation platform that launched a dedicated Agents product in February 2026 and repositioned from building workflows to building a team of AI agents. What makes it worth attention is not the pivot but the constituency it has chosen: operations and compliance teams that want AI automation and require a human approval gate before anything fires.

Human in the loop is treated as a first class primitive rather than a feature bolted on later. Approvals, assigned tasks and human data entry can be placed at any critical stage of a workflow, and in flow approvals combine with branching Paths so a process can adapt on the outcome of a human decision. That is a materially different design centre from platforms that treat oversight as an exception path, and it is the reason the platform lands with finance, support and operations teams running repeatable multi step processes.

Models are included rather than assembled. Built in access covers GPT-5.1, Claude 4.5 Opus and Gemini 3 Pro without requiring external API keys, so a buyer is not managing separate provider accounts. Native Tables, added in November 2025, provide a place to store, query and manage data inside the platform, and workflows can be edited conversationally. Around 200 app integrations are included on every plan rather than gated by tier, which removes a common upgrade trap, though the library remains far smaller than Zapier's or Make's and niche systems typically need webhooks.

Security is documented: SOC 2 Type II and ISO/IEC 27001:2022. The notable gap is healthcare. There is no HIPAA compliance or business associate agreement, which a reviewer identified as the specific blocker preventing adoption for some workflows despite an otherwise solid foundation. Two further limits are worth pricing in: AI credits deplete quickly on high volume or complex agent runs, and when an AI output does not match an expected schema the workflow can stumble, with error handling for that case still thin.

Vendor details

Canonical URL

https://www.relay.app

Company status

independent

Use cases & customers

Target customers

operations and compliance teamsmid market SaaS, agencies and service firmssupport, sales and finance teams

Deployment options

SaaScloud

In practice

Your compliance team will not allow an agent to act without sign off. Approvals, tasks and human data entry sit at any critical stage rather than as an exception path, and branching logic can react to the decision.

You do not want to manage OpenAI, Anthropic and Google API keys separately. GPT-5.1, Claude 4.5 Opus and Gemini 3 Pro are built in without external keys.

You keep hitting upgrade walls to unlock the integration you actually need. Relay includes all integrations on every plan and charges on users, steps and AI credits instead.

Sources & related URLs

Research sources

https://www.stork.ai/en/relay-app https://toarn.com/insights/relay-app https://www.lindy.ai/blog/relay-app-features https://www.producthunt.com/products/relay-app

Agentic Index coverage score

10.0 / 14 capabilities · 71%

Integrations & Tool CallingAround 200 app integrations including Microsoft Planner, SharePoint, Notion, Gmail, Slack and HubSpot, with all integrations included on every plan rather than gated by tier. Sources conflict on the count, reporting both 100-150 and 200+, and the library is materially smaller than Zapier or Make, with niche systems requiring webhooks. 2026-08-05	Full
Workflow OrchestrationMulti step workflows combining AI actions, conditional branching Paths and in flow approvals, with a dedicated Agents product launched February 2026 adding agent building on top of the workflow engine. 2026-08-05	Full
Knowledge Grounding & RAGBuilt in AI actions cover summarisation, translation, data extraction and classification, and Native Tables hold structured data the workflow can query, but there is no retrieval layer over a document corpus, vector store or citation mechanism. 2026-08-05	Partial
Human Oversight & GuardrailsThe clearest case of human oversight as a design centre rather than an exception path in this index. Approvals, assigned tasks and human data entry can sit at any critical workflow stage, in flow approvals combine with branching logic so the process adapts to the decision, and independent analysis identifies the target buyer explicitly as operations and compliance teams needing approval gates before anything fires. 2026-08-05	Full
Security, Identity & GovernanceSOC 2 Type II and ISO/IEC 27001:2022 compliance, with a SOC 2 link published in the site footer alongside an information security page. NOTABLE GAP: no HIPAA compliance and no business associate agreement, identified by a reviewer as the specific blocker preventing use for healthcare adjacent workflows. 2026-08-05	Full
Observability & AuditabilityAn audit trail is cited as an enterprise relevant capability and a public system status page exists, and approval steps inherently record who authorised what, but no run tracing, agent decision transparency or execution analytics is documented. 2026-08-05	Partial
Memory & State PersistenceNative Tables, launched November 2025, provide first class storage for holding, querying and managing data inside the platform across runs, so workflows and agents carry state rather than passing it through each execution. 2026-08-05	Full
Deployment & Data ResidencyMulti tenant SaaS only. No self hosted, VPC, private cloud or regional data residency option is documented, which is a real constraint for the compliance focused buyer the product otherwise targets. 2026-08-05	Unable to verify
Prebuilt Agents, Templates & PacksShips prebuilt AI actions for summarisation, translation, extraction and classification, plus a no code agent builder with prebuilt capabilities aimed at beginners and a drag and drop visual builder praised by non technical users. 2026-08-05	Full
Triggers & Channel CoverageCustomisable workflow triggers initiate automations on specific events across the connected app library, with webhooks covering systems outside it, and human assigned tasks acting as a further entry point into a running workflow. 2026-08-05	Full
Model Flexibility & RoutingNative access to GPT-5.1 from OpenAI, Claude 4.5 Opus from Anthropic and Gemini 3 Pro from Google built into the platform WITHOUT requiring external API keys, so model choice sits with the builder and no separate provider accounts are needed. 2026-08-05	Full
APIs, SDKs & MCP ExtensibilityWebhooks extend the platform to systems outside the integration library, which reviewers note is the standard workaround for niche software, and documentation exists, but no public API, SDK or MCP surface was retrieved. 2026-08-05	Partial
Testing, Debugging & OptimizationChat based workflow editing and the visual builder support iteration, and human approval steps catch bad output before it acts, but a reviewer identifies error handling as a specific weakness: when an AI output does not match the expected schema the workflow can stumble, and no evaluation harness or schema validation tooling is documented. 2026-08-05	Partial
Browser & Computer UseNo browser control or computer use capability. Agents act through app integrations and webhooks rather than by operating interfaces. 2026-08-05	Unable to verify

The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

Free tier available; Team plan reported at $138 per month for 10 users; usage beyond included runs at about $0.05 per automation run; 50 percent annual discount

users, workflow steps and AI credits, kept separate

Free tierTrial available

What is public

Plan structure, the all integrations included policy and the free tier are public; specific rates circulate mainly through third party analyses.

Billing mechanics

Tiered on three separate axes: users, workflow steps and AI credits, deliberately kept distinct so buyers can see what they are paying for. All integrations are included on every tier rather than gated by plan. A Team plan is reported at $138 a month for ten users, a 50 percent annual discount anchors longer commitments, and usage beyond included runs is charged at roughly $0.05 per automation run.

Cost watchouts

AI credits are the variable that bites: independent reviews report they deplete quickly on high volume or complex agent runs, leading to top up spend that the per user headline does not suggest

Variable cost rationale

Per user pricing is predictable but AI credits are metered and reported to deplete quickly on complex agent runs, so the agentic half of the product carries the variable exposure while the workflow half does not.

Additional watchouts

Two costs sit outside the headline. AI credits exhaust quickly on complex agent runs, and the platform has no HIPAA compliance or business associate agreement, so healthcare adjacent workflows are blocked regardless of budget. A buyer should also weigh the smaller integration library, since anything outside roughly 200 apps needs webhook work.

Overage / add-ons

Approximately $0.05 per automation run beyond free tier limits or plan included runs, with AI credits metered separately from workflow steps

Sales call required

Mixed (some tiers require a call)

Free / trial

Free tier with limited runs, sufficient to build and test real workflows before paying

Lowest paid plan

Team

Commercial notes

The pricing is a deliberate attack on Zapier's per task billing, which independent analysis notes can spiral at volume, and including every integration on every plan removes the upgrade trap that competitors rely on. Set against that, funding of $8.1 million is flagged by the same analysis as limiting enterprise sales motion and support scale, which matters for a product now selling to compliance conscious operations teams.

Key ambiguities

Rates come from third party analyses rather than a retrieved first party rate card, and the relationship between included steps, included AI credits and the $0.05 per run overage is not fully specified. Verify at relay.app/pricing before quoting.

Missing data

First party confirmation of tier rates, included step and credit allowances per tier, and enterprise pricing.

Agentic Index verified 2026-08-05

Official pricing page

Related vendors

Alternatives to Relay.app

The closest documented capability profiles to Relay.app among agent builders tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

Taskade9.0 / 14Adds documented Browser & Computer Use
Teneo10.0 / 14Adds documented Deployment & Data Residency
Base4410.5 / 14Adds documented Deployment & Data Residency
Gumloop11.5 / 14Adds documented Deployment & Data Residency
AgentX10.0 / 14Adds documented Deployment & Data Residency
Airia10.0 / 14Adds documented Deployment & Data Residency

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Compare Relay.app with other vendors →Compare all agent builders →See where Relay.app ranks →

Data confidence: high