How much does Kong cost?

Not retrieved in this pass. Kong runs an open source core with commercial Konnect and enterprise tiers, and separately ships Konnect Metering and Billing so customers can meter and monetise their own AI consumption, but no rate card for Kong itself was obtained.

What are the best alternatives to Kong?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Kong among agent infrastructure platforms are Metorial, Arcade, Stacklok, nexos.ai. Each is graded on the same public evidence, and no vendor pays for placement.

Does Kong support MCP and tool calling?

Full under the Agentic Index verification standard. Kong publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: Native first class Model Context Protocol support described as the first among major gateways, plus A2A protocol support, an extensive plugin architecture, Insomnia for API and MCP development, a developer portal, and generation of SDKs and. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Kong have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. Kong 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: Security governance IS the product: authentication and authorisation through OpenID Connect, JWT and access control lists, token and request level rate limiting, prompt security, content moderation and PII sanitisation, and critically IDENT. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can Kong be self hosted or run in a private VPC?

Full under the Agentic Index verification standard. Kong publishes public evidence of self hosted or private VPC deployment, along with data residency options for controlling where agent data is processed and stored. Evidence basis: Runs as a self hosted data plane with the Konnect managed control plane, in hybrid configurations or fully on premise, and Kong explicitly frames perimeter enforcement as keeping data within the customer's own sovereign boundaries. A publis. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Kong

Q: Who is Kong for?

large enterprise, platform and API teams, security and governance teams, regulated industries with sovereignty requirements

Also known as: Kong Inc, Kong AI Gateway, Kong Agent Gateway, Kong Konnect, Kong Gateway, Kong MCP Registry, Kong Insomnia, KAi

Visit site

Agent infrastructureindependentVerified 2026-08-07

Entry priceOpen source core free; Konnect and enterprise tiers not retrievedFull pricing detail

API gateway extended into agent infrastructure: Kong Agent Gateway governs LLM, MCP and agent to agent traffic from one control plane, with identity verified at every hop and policy written against request meaning.

Kong built its business on API gateways and has extended that position into what it calls AI Connectivity, a unified governance and runtime layer across APIs, events, LLM calls, MCP connections and agent to agent communication. The claim it makes is architectural: every request, tool call and agent interaction flows through one control plane, so visibility, policy and audit are consistent across the whole agentic stack rather than stitched together from point solutions.

The agentic piece landed in April 2026 with Kong Agent Gateway inside AI Gateway 3.14, which extends governance to agent to agent traffic over the emerging A2A protocol. Kong positions this as covering all three AI protocol types, LLM, MCP and A2A, on one platform, one runtime and one control plane, and argues that teams adopting a standalone MCP broker will have to integrate it with LLM and agent governance later anyway.

Underneath sit concrete controls. Authentication and authorisation through OpenID Connect, JWT and access control lists. Rate limiting on both requests and token consumption to cap cost. Guardrails covering prompt security, content moderation and personally identifiable information sanitisation, enforced at the network perimeter rather than relying on the model provider. A semantic policy engine using on gateway vector embeddings so policy can be written against the meaning of a request rather than its shape. And identity verified at every hop of a multi agent chain, so when an agent calls an MCP tool that triggers an API that delegates to a second agent, each component authenticates independently.

The honest limits are that this is a control plane, not an agent platform. Kong does not author or orchestrate agent workflows, and no agent memory or computer use capability exists. Its own corporate compliance certifications were not retrieved in this pass.

Vendor details

Canonical URL

https://konghq.com

Subcategory

AI and agent traffic control plane

Funding status

Private, San Francisco. Long established in API management before extending into AI connectivity. Funding detail not retrieved in this pass.

Company status

independent

Use cases & customers

Primary use cases

governing LLM, MCP and agent to agent traffic from one control planesecuring and discovering MCP servers used by AI agentsAI cost control through token level rate limiting and meteringenforcing guardrails and PII redaction at the network perimeterAPI management extended to agentic workloads

Target customers

large enterpriseplatform and API teamssecurity and governance teamsregulated industries with sovereignty requirements

Deployment options

Kong Konnect managed control planeself-hostedhybridon-premise within sovereign boundaries

Integrations

An API gateway heritage means very broad backend integration, extended to AI through LLM provider routing with load balancing and failover, native Model Context Protocol support, and an MCP Registry inside the Konnect API Service Catalog for registering, discovering and governing the MCP servers and tools agents use. Kong can generate MCP servers from existing API specifications, and Insomnia supports MCP testing.

In practice

A platform team registers every internal MCP server in one catalogue so agents discover governed tools instead of engineers spinning up unmanaged servers across the business.

A security team enforces PII redaction and prompt guards at the gateway rather than trusting each model provider, keeping regulated data inside its own boundary.

A finance team caps token consumption per team at the gateway and meters AI usage back to internal cost centres rather than discovering the bill at the end of the month.

Sources & related URLs

Research sources

https://konghq.com/products/kong-ai-gateway https://www.prnewswire.com/news-releases/kong-ai-gateway-now-supports-agent-to-agent-traffic-becoming-the-most-comprehensive-ai-gateway-for-the-agentic-era-302741741.html https://www.prnewswire.com/news-releases/kong-to-unveil-ai-connectivity-vision-and-roadmap-to-power-the-agentic-era-302676445.html https://konghq.com/blog/product-releases/enterprise-mcp-gateway https://konghq.com/blog/learning-center/what-is-a-mcp-gateway https://konghq.com/blog/enterprise/api-gateway-governance

Agentic Index coverage score

9.5 / 14 capabilities · 68%

Integrations & Tool CallingAPI gateway heritage gives very broad backend reach, extended to AI with LLM provider routing, native Model Context Protocol support, an MCP Registry in the Konnect API Service Catalog for registering and discovering agent tools, and the ability to generate MCP servers from existing API specifications. Kong AI Gateway product page and MCP Gateway release 2026-08-07	Full
Workflow OrchestrationRoutes and governs multi agent traffic including agent to agent communication, with dynamic routing of tool calls to the right backend plus load balancing and failover, but Kong is a CONTROL PLANE not an agent platform: it does not author or orchestrate agent workflows. Kong AI Gateway and Agent Gateway documentation 2026-08-07	Partial
Knowledge Grounding & RAGA semantic policy engine uses lightweight on gateway vector embeddings to interpret request meaning, and the MCP Registry catalogues the tools and resources agents can reach, but Kong does not ground agents in customer knowledge and ships no retrieval layer of its own. Kong AI connectivity roadmap analysis 2026-08-07	Partial
Human Oversight & GuardrailsGuardrails are enforced at the network perimeter rather than left to the model provider, covering prompt security, content moderation and PII sanitisation, with policy enforceable on request meaning; but these are AUTOMATED controls and no human approval gate, review step or escalation path is documented. Kong AI Gateway governance documentation 2026-08-07	Partial
Security, Identity & GovernanceSecurity governance IS the product: authentication and authorisation through OpenID Connect, JWT and access control lists, token and request level rate limiting, prompt security, content moderation and PII sanitisation, and critically IDENTITY VERIFIED AT EVERY HOP of a multi agent chain so each component authenticates independently rather than inheriting the initiating identity. Guardrails run at the perimeter keeping data inside sovereign boundaries. VERIFICATION FLAG: Kong's own corporate attestations were not retrieved in this pass. Kong AI Gateway and gateway governance documentation 2026-08-07	Full
Observability & AuditabilityUnified visibility and observability across ALL AI traffic types including LLM, MCP and agent to agent, with auditability described as baked in, enterprise wide observability of the full AI data path, and Konnect Metering and Billing giving token level consumption reporting. Kong Agent Gateway release and AI Connectivity roadmap 2026-08-07	Full
Memory & State PersistenceNo agent memory or conversation state persistence is documented; the gateway governs traffic in flight rather than holding agent state. Honest absence, and architecturally consistent with what a control plane does. Kong product documentation 2026-08-07	Unable to verify
Deployment & Data ResidencyRuns as a self hosted data plane with the Konnect managed control plane, in hybrid configurations or fully on premise, and Kong explicitly frames perimeter enforcement as keeping data within the customer's own sovereign boundaries. A published regional residency matrix was not retrieved. Kong Konnect documentation and roadmap analysis 2026-08-07	Full
Prebuilt Agents, Templates & PacksA large plugin ecosystem supplies ready made policy components including PII redaction, prompt guards, rate limiting and auth, the MCP Registry ships as a catalogue capability, and Kong can generate MCP servers, TypeScript SDKs and Python clients from an existing API specification. Kong product pages and engineering blog 2026-08-07	Full
Triggers & Channel CoverageAll AI traffic flows through the gateway and real time streaming is supported, so coverage of the request path is total, but the model is request driven interception rather than an event, schedule or channel trigger surface. Kong AI Gateway documentation 2026-08-07	Partial
Model Flexibility & RoutingCore function: govern how developers, applications and agents consume LLMs, routing prompts across providers with load balancing and failover, controlling access, data leakage and token usage, with cost governance and monetisation through Konnect Metering and Billing. Model routing is the product rather than a feature. Kong AI Gateway product page 2026-08-07	Full
APIs, SDKs & MCP ExtensibilityNative first class Model Context Protocol support described as the first among major gateways, plus A2A protocol support, an extensive plugin architecture, Insomnia for API and MCP development, a developer portal, and generation of SDKs and MCP servers from API specifications. Kong Agent Gateway release and developer documentation 2026-08-07	Full
Testing, Debugging & OptimizationMCP testing is supported in Kong Insomnia, giving a pre production validation path for agent tool integrations, but no agent evaluation harness, scoring or regression suite is documented. Kong AI Connectivity roadmap release 2026-08-07	Partial
Browser & Computer UseKong operates on network traffic between agents, tools and models; no browser control, page navigation or computer use capability exists or would be architecturally expected. Kong product documentation 2026-08-07	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

Open source core free; Konnect and enterprise tiers not retrieved

not retrieved

Free tierTrial available

Included quota

Not retrieved.

What is public

Product structure and the open core model are public. No prices were retrieved in this pass.

Billing mechanics

Not retrieved in this pass. Historically an open source data plane with paid Konnect control plane and enterprise licensing.

Cost watchouts

This is a classic open core split and the agentic capabilities sit on the commercial side: AI Gateway, Agent Gateway, the MCP Registry, semantic policy and Konnect governance are not the free gateway. Because Kong sits in the request path for all AI traffic, its own cost scales with the volume of agent and model traffic an organisation runs, which is the same quantity the customer is trying to control by adopting it. Konnect Metering and Billing is a separate product for metering the customer's OWN AI consumption, not a discount on Kong.

Variable cost rationale

Graded medium rather than high because the open source core gives a genuinely free and self hostable entry point and the commercial model follows established API management norms rather than an invented agent consumption unit. The offsetting risks are that Kong sits in the path of all AI traffic so its cost tracks the very volume being governed, and that the tier boundary around agentic governance features was not retrieved.

Additional watchouts

Establish the tier boundary around the agentic capabilities early. The free gateway is not the AI Gateway, and the governance features that justify the purchase are the ones most likely to sit in the paid tier.

Overage / add-ons

Not retrieved.

Sales call required

Mixed (some tiers require a call)

Free / trial

Open source Kong Gateway is free; Konnect offers a free tier historically

Commercial notes

Kong is a long established API management vendor extending into AI connectivity, so it arrives with an existing enterprise commercial motion rather than a new one. Notably it also sells the tooling for customers to meter and monetise their own AI consumption.

Key ambiguities

All commercial terms, and specifically which of the AI and agent capabilities are gated to which tier. Given the open core security paywall pattern seen elsewhere in this index, the tier boundary around governance features is the thing worth establishing before purchase.

Missing data

Every commercial detail: Konnect tier pricing, enterprise licensing, and which AI and agent capabilities are gated behind which tier.

Agentic Index verified 2026-08-07

Official pricing page

Related vendors

Alternatives to Kong

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

Metorial8.5 / 14A lighter documented profile than Kong
Arcade8.0 / 14A lighter documented profile than Kong
Stacklok11.0 / 14Fuller documented coverage on Workflow Orchestration and Knowledge Grounding & RAG
nexos.ai9.5 / 14Adds documented Memory & State Persistence
Paragon11.5 / 14Adds documented Memory & State Persistence
AIsa7.0 / 14A lighter documented profile than Kong

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 Kong with other vendors →Compare all agent infrastructure platforms →See where Kong ranks →

Data confidence: high