Skip to content

Blocks.ai publishes this comparison and sells in this market. Every claim about another product links to a public source.

Kong AI Gateway vs LangSmith

Kong AI Gateway

Gateway that governs LLM, MCP, and agent-to-agent traffic

LangSmith

Platform for observing, evaluating, and deploying agents

Short answer

Kong AI Gateway is a gateway for LLM, MCP, and A2A traffic; LangChain calls LangSmith its platform for observing, evaluating, and deploying agents.⁠Source 1, Source 2, Source 3 Kong proxies to agents registered as upstream URLs and can apply per-agent allow or deny lists; LangSmith Deployment runs agents in LangChain’s cloud (Plus or above) or, on Enterprise, self-hosted or hybrid.⁠Source 4, Source 5, Source 6, Source 7

Where each one sits

Six layers of running AI agents at a company, and what each product’s own public sources say it covers.

These aren’t the same kind of product

Where they overlap
Both can authenticate calls to an agent, and both handle A2A and MCP: Kong proxies them, and LangSmith’s Agent Server serves them.⁠Source 2, Source 4, Source 8, Source 9, Source 10
Where they differ
LangSmith builds agents without code, in Fleet, and runs agents on LangSmith Deployment.⁠Source 5, Source 11 Kong proxies to agents registered as upstream URLs; agent-building tools in Kong AI Gateway aren’t publicly documented.⁠Source 4
Running both
Kong says a LangGraph agent can send its model calls through Kong AI Gateway, and its walkthrough monitors the agent’s LangGraph Server with LangSmith.⁠Source 12
Public sources · checked 2 October 2026
  • Offered
  • Preview
  • Not publicly documented

Kong AI Gateway

  • Build agents: Not publicly documented
  • Host and run agents: Not publicly documented
  • Identity and access: OfferedPer-agent allow or deny lists⁠Source 4
  • Registry and governance: PreviewKonnect Catalog agents⁠Source 13
  • Traffic between agents, tools, and models: OfferedGateway for LLM, MCP, A2A⁠Source 1
  • Agents across organizations: Not publicly documented

LangSmith

  • Build agents: OfferedFleet agent builder⁠Source 11
  • Host and run agents: OfferedLangSmith Deployment runtime⁠Source 5
  • Identity and access: OfferedFleet per-agent permissions⁠Source 14
  • Registry and governance: OfferedCentralized registry in Deployment⁠Source 15
  • Traffic between agents, tools, and models: OfferedFleet forwards MCP tool calls⁠Source 16
  • Agents across organizations: Not publicly documented

At a glance

TopicKong AI GatewayLangSmith
What it isA gateway for LLM, MCP, and A2A traffic, with shared authentication, observability, and policy features.⁠Source 1, Source 2LangChain calls it a platform for observing, evaluating, and deploying agents, with Fleet for creating agents without code.⁠Source 3, Source 11
Where agents runAt an upstream URL registered for each agent, which the gateway proxies to.⁠Source 4 In 2.x, data plane nodes you run carry the traffic.⁠Source 2On LangSmith Deployment, in LangChain’s cloud (Plus plan or above) or, on Enterprise, self-hosted or hybrid.⁠Source 5, Source 6, Source 7 Agents hosted elsewhere can send traces.⁠Source 6
Access controlThe gateway can require an API key or OpenID Connect login, then check a per-agent allow or deny list.⁠Source 4, Source 17Hosted agents take LangSmith API keys by default, or your own auth handlers.⁠Source 8 Fleet has per-agent Clone, Run, and Edit permissions.⁠Source 14, Source 18
PricingAI Management Plus from $25 a month plus usage; control planes are $200 a month hybrid, $500 Dedicated Cloud, or $25 serverless.⁠Source 19 Enterprise is custom, billed annually.⁠Source 19Developer has one free seat; Plus is $39 per seat a month; Enterprise is through sales.⁠Source 7 Usage is metered in LSUs at $1 each.⁠Source 7
Generally availableKong announced version 2.0 as generally available on 1 September 2026.⁠Source 20 Konnect Catalog’s agent inventory is in beta.⁠Source 13LangChain announced GA of Deployment, as LangGraph Platform, on 14 May 2025, and of Fleet, as Agent Builder, on 13 January 2026.⁠Source 15, Source 21, Source 22, Source 23 Self-hosting Fleet is in beta.⁠Source 11

What each one is

Kong AI Gateway

Kong AI Gateway is one gateway for LLM, MCP, and agent-to-agent (A2A) traffic, with shared authentication, observability, and policy features.⁠Source 1, Source 2 In 2.x, an agent is added as an AI Agent entity, which exposes it at a gateway endpoint and can carry policies.⁠Source 4, Source 24

LangSmith

LangChain describes LangSmith as a framework-agnostic platform for observing, evaluating, and deploying agents.⁠Source 3 LangSmith Deployment runs agents in production, and LangSmith Fleet creates and manages agents without code; LangChain positions Fleet for non-technical teams.⁠Source 3, Source 5, Source 11 Agents hosted elsewhere can send traces to LangSmith.⁠Source 6

The differences that matter

  1. Who can call an agent

    Kong AI Gateway

    The gateway can require an API key or OpenID Connect login, such as Okta or Azure AD, then check a per-agent allow or deny list.⁠Source 4, Source 17

    LangSmith

    Hosted agents use LangSmith API keys by default, or your own auth handlers; Fleet has per-agent Clone, Run, and Edit permissions.⁠Source 8, Source 14

    By default, Kong adds no credentials when it forwards a request to the agent.⁠Source 4 Self-hosted Agent Servers have no default authentication.⁠Source 8

  2. Model and tool access

    Kong AI Gateway

    Kong gives one API to major LLM providers and can apply AI Policies to models and MCP servers; MCP bundling shows callers only authorized tools.⁠Source 2, Source 20, Source 25

    LangSmith

    Fleet uses managed model tiers needing no provider key of your own; on Enterprise, admins can control who uses which tools and MCP servers.⁠Source 18, Source 26

    With Kong, you supply your own provider credentials.⁠Source 2 LangChain says the model behind each Fleet tier may change over time.⁠Source 26

  3. What gets recorded

    Kong AI Gateway

    A2A audit logs record task IDs, method calls, latencies, and errors, and A2A telemetry can flow to Konnect and OpenTelemetry.⁠Source 4, Source 27

    LangSmith

    Each deployment gets a tracing project, Fleet traces include tool calls and outputs, and LangSmith accepts traces from any OpenTelemetry-compatible app.⁠Source 18, Source 28, Source 29

    Konnect telemetry leaves out request and response bodies unless you opt in.⁠Source 2 LangSmith keeps base traces 14 days and extended traces 180 days.⁠Source 7

For security teams

What a security review asks, answered from each vendor’s public documentation.

TopicKong AI GatewayLangSmith
Network exposureIn 2.x, data plane nodes you run proxy to each agent’s URL and authenticate to the control plane over mTLS.⁠Source 2, Source 4Enterprise customers can reach LangSmith Cloud over AWS PrivateLink or GCP Private Service Connect, off the public internet.⁠Source 30
IdentityCan require an API key or OpenID Connect (Okta, Azure AD, or another provider) before routing.⁠Source 4, Source 17 Enterprise adds Konnect SSO.⁠Source 19, Source 31Hosted Agent Servers default to API keys; self-hosted ones have no default authentication.⁠Source 8 Cloud Enterprise adds SAML SSO and SCIM.⁠Source 32
Access changes and revocationEach agent has an enabled switch; Request Termination or deny lists block callers.⁠Source 4, Source 33, Source 34 Nodes keep their last config without Konnect.⁠Source 2Fleet sharing can be revoked at any time.⁠Source 18 Admins can deactivate any member’s access token, effective within a minute.⁠Source 35
Audit trailA2A audit logs: task IDs, method calls, latencies, errors.⁠Source 27 Konnect audit logs (Enterprise): webhook delivery, kept 7 days in Konnect.⁠Source 19, Source 36Enterprise audit logs of admin actions, kept up to 400 days; an hourly job can forward them to a SIEM.⁠Source 37
Compliance2.2+ FIPS mode: FIPS 140-3 algorithms only, not NIST-validated.⁠Source 38 Kong Inc. lists ISO 27001 and SOC 2 (report under NDA).⁠Source 39LangChain’s Trust Center lists SOC 2 Type II, ISO 27001:2022, GDPR, and HIPAA for LangSmith.⁠Source 40

Full comparison

18 criteria in five groups. Every cell links to its source, or says no public source answers it.

Kong AI Gateway and LangSmith compared on 18 criteria
Kong AI GatewayLangSmith
What it is
What it is and who it’s forOne gateway that governs LLM, MCP, and A2A traffic.⁠Source 1, Source 2 All three run through one data plane, with shared authentication, observability, and policy features.⁠Source 2LangChain calls it a framework-agnostic platform for observing, evaluating, and deploying agents.⁠Source 3 It positions its no-code part, Fleet, for non-technical teams.⁠Source 3, Source 11
MaturityVersion 2.0 announced generally available on 1 September 2026; 2.2.0 released 30 September 2026.⁠Source 20, Source 41 Konnect Catalog’s agent inventory is in beta.⁠Source 13LangChain announced GA of LangGraph Platform (now Deployment) on 14 May 2025 and Agent Builder (now Fleet) on 13 January 2026.⁠Source 15, Source 21, Source 22, Source 23
Control
Agent registry and discoveryIn 2.x, each AI Agent entity, A2A or plain HTTP, is scoped to one gateway instance.⁠Source 2, Source 4 Konnect Catalog’s agent inventory is in beta.⁠Source 13LangChain says Deployment manages agents in a centralized registry with versioning and rollbacks.⁠Source 15, Source 42 Listing agents hosted elsewhere is not publicly documented.
Identity and access controlCan require an API key or OpenID Connect login, such as Okta or Azure AD, before routing, and check a per-agent allow or deny list.⁠Source 4, Source 17Hosted agents use LangSmith API keys by default, or your auth handlers.⁠Source 8 Fleet has per-agent Clone, Run, and Edit permissions.⁠Source 14 RBAC is Enterprise.⁠Source 43
Ownership, policy, and revocationAgent policies include input validation, logging, and rate limits.⁠Source 4 Kong’s AI PII Sanitizer scrubs requests to AI models: a Plus add-on, included on Enterprise.⁠Source 19, Source 44Fleet agent cards show the owner, sharing can be revoked at any time, and pausing an agent’s channels pauses it.⁠Source 18, Source 23, Source 45 Admins can set spend limits.⁠Source 46
Audit log and observabilityA2A audit logs record task IDs, method calls, latencies, and errors.⁠Source 27 Konnect platform audit logs are Enterprise-only, sent to a webhook, and kept 7 days.⁠Source 19, Source 36Enterprise audit logs of admin and configuration actions, kept up to 400 days.⁠Source 37 Each deployment gets a tracing project.⁠Source 28
Connection
How agents connectEach agent is an upstream URL the gateway proxies to.⁠Source 4 In 2.x, data plane nodes you run forward allowed traffic and stay connected to Konnect.⁠Source 2On self-hosted LangSmith with Deployment, agents run behind an ingress you configure.⁠Source 47 Enterprise can reach LangSmith Cloud over AWS PrivateLink.⁠Source 30
Agents across organizationsNot publicly documented (checked 2 October 2026)Agent Server agents can talk to other A2A-compatible agents.⁠Source 9 An access model for another organization’s agents is not publicly documented.
Protocol supportDetects A2A over JSON-RPC and REST bindings.⁠Source 4 Accepts four MCP revisions, 2025-03-26 to 2026-07-28 (the last from version 2.1).⁠Source 48Agent Server supports A2A over JSON-RPC only, with documented gaps, and can expose agents as MCP tools.⁠Source 9, Source 10 Fleet discovers tools from remote MCP servers.⁠Source 16
Frameworks, models, and clouds supportedKong says it governs A2A traffic without changing how agents are built.⁠Source 49 Agents that don’t use A2A can pass through as plain HTTP.⁠Source 4Deployment runs Google ADK, Strands, CrewAI, and AutoGen agents via a wrapper or the LangGraph Functional API.⁠Source 50 Agents hosted elsewhere can send traces.⁠Source 6
Operations
Deployment options and data residency2.x: a Konnect-managed control plane in a region you choose, data plane nodes you run.⁠Source 2, Source 51 Kong also sells Kong-managed Dedicated Cloud and serverless gateways.⁠Source 19, Source 52, Source 53Cloud on GCP (US, EU, APAC) or AWS (US).⁠Source 6, Source 30 On Enterprise: BYOC (GA on AWS), self-hosted, or hybrid.⁠Source 6, Source 7, Source 54
Compliance attestationsFrom 2.2, FIPS mode uses only FIPS 140-3 approved algorithms; not submitted for NIST validation.⁠Source 38 Kong lists ISO 27001, SOC 2, and PCI DSS for Kong Inc.⁠Source 39LangChain’s Trust Center lists SOC 2 Type II, ISO 27001:2022, GDPR, and HIPAA, with a 2026 SOC 2 Type II report and ISO 27001 certificate.⁠Source 40
Support and SLAKonnect targets 99.9% availability; Dedicated Cloud Gateways list a 99.99% SLA, serverless gateways none.⁠Source 19 Enterprise support SLAs: 30 min to 2 hours.⁠Source 19Base, Standard, and Premium support plans.⁠Source 55 Reasonable commercial efforts toward 99.5% API uptime for SaaS, measured quarterly.⁠Source 55
Time and effort to get runningA quickstart script creates a Konnect control plane and a local Docker data plane.⁠Source 25 To route A2A traffic, create an AI Agent entity and attach policies.⁠Source 24Cloud deployment needs the Plus plan or above and an API key.⁠Source 56 Self-hosting the Deployment control plane needs Kubernetes, self-hosted LangSmith, and Enterprise.⁠Source 5
Pricing model and public pricesPlus: from $25 a month plus usage; per control plane, $200 hybrid, $500 Dedicated Cloud, $25 serverless.⁠Source 19 Enterprise: custom, billed annually.⁠Source 19Developer: one free seat.⁠Source 7 Plus: $39 per seat a month, with one free small serverless deployment.⁠Source 7 Enterprise: contact sales.⁠Source 7 Usage is metered in LSUs, $1 each.⁠Source 7
Building
Agent building toolsKong says it can generate MCP tools and servers from Kong-managed APIs.⁠Source 1 Tools for building agents in Kong AI Gateway are not publicly documented.Fleet builds agents from a description or a template, without code.⁠Source 11 Code-first agents can be deployed with the LangGraph CLI and app templates.⁠Source 5
Model accessOne API to providers including OpenAI, Anthropic, Amazon Bedrock, Gemini, Azure AI, Mistral, and Ollama, with credentials you supply.⁠Source 2, Source 25, Source 57Fleet’s managed tiers (Fast, Pro, Max) need no model-provider key of your own; custom models aren’t in that picker.⁠Source 26 LangChain says closed or open models work.⁠Source 3
Integrations and ecosystemA Policies Hub of policies and integrations.⁠Source 33 AI Vault works with AWS, GCP, Azure, and HashiCorp Vault secrets backends.⁠Source 2Fleet has built-in tools such as Gmail, Slack, and GitHub; LangChain says Fleet agents work in Slack, Teams, and Gmail.⁠Source 26, Source 58

Which to choose

Choose Kong AI Gateway if

  • You want one gateway for LLM, MCP, and A2A traffic, with shared authentication, observability, and policy features.⁠Source 1, Source 2
  • You want to govern agents at their existing URLs; Kong says it governs A2A traffic without changing how agents are built.⁠Source 4, Source 49
  • You want callers authenticated by API key or your OpenID Connect provider, with per-agent allow or deny lists and rate limits.⁠Source 4, Source 17
  • You want one API to model providers such as OpenAI, Anthropic, Gemini, and Amazon Bedrock, switching or combining them without rewriting integrations.⁠Source 25, Source 57

Choose LangSmith if

  • You want a platform that builds and runs agents: Fleet builds them without code, and LangSmith Deployment runs agents in production.⁠Source 5, Source 11
  • You want to trace agents wherever they run, from any OpenTelemetry-compatible app or through integrations such as OpenAI, Anthropic, and CrewAI.⁠Source 6, Source 29, Source 59
  • You need LangSmith’s control plane and Agent Servers in your own infrastructure, for full data residency or air-gapped use, on Enterprise.⁠Source 5
  • You want Fleet’s per-agent Clone, Run, and Edit permissions, and human approval before specific agent actions.⁠Source 14, Source 18

Questions buyers ask

Is Kong AI Gateway an alternative to LangSmith?

Partly: both can authenticate calls to agents and handle A2A and MCP.⁠Source 2, Source 4, Source 8, Source 9, Source 10 LangSmith also builds, runs, and traces agents; Kong AI Gateway proxies to agents registered as upstream URLs and routes LLM traffic to major providers through one API.⁠Source 4, Source 5, Source 11, Source 25, Source 28

Do Kong AI Gateway and LangSmith support MCP and A2A?

Kong detects A2A over JSON-RPC and REST bindings and accepts four MCP revisions, the newest from version 2.1.⁠Source 4, Source 48 LangSmith’s Agent Server serves A2A over JSON-RPC only, with documented gaps, and can expose agents as MCP tools; Fleet uses tools from remote MCP servers.⁠Source 9, Source 10, Source 16

Can either one be self-hosted?

Fully self-hosted Kong AI gateways are part of Kong’s separate Gateway Enterprise offering; the 2.x AI entities use a Konnect-managed control plane.⁠Source 2, Source 19 On LangSmith’s Enterprise plan, LangSmith Deployment can run in your own Kubernetes cluster, including air-gapped, and LangSmith comes as BYOC on AWS.⁠Source 5, Source 6, Source 54

How is each one priced?

Kong’s AI Management Plus starts at $25 a month plus usage, with $200 a month per hybrid control plane; Enterprise is custom.⁠Source 19 LangSmith has one free Developer seat; Plus is $39 per seat a month; Enterprise is through sales; usage is metered in $1 units.⁠Source 7

How we compare

Read the full method

Every claim on this page links to a public source. Where none answers a question, the page says so.

We re-check every fact at least every 90 days. This page was last checked .

Something wrong or out of date? Tell us and we’ll correct it.

Sources

66 public sources, each with the date we checked it. Every one opens in a new tab.

  1. Source 1: Kong AI Gateway product page Kong · checked Back:abcdefg

  2. Source 2: AI Gateway architecture - Kong Docs Kong · checked Back:abcdefghijklmnopqrst

  3. Source 3: LangSmith: The Agent Engineering Platform LangChain · checked Back:abcdefg

  4. Source 4: AI Agents - Kong AI Gateway docs Kong · checked Back:abcdefghijklmnopqrstuvwx

  5. Source 5: LangSmith Deployment LangChain · checked Back:abcdefghijk

  6. Source 6: Set up LangSmith LangChain · checked Back:abcdefghi

  7. Source 7: LangSmith Plans and Pricing LangChain · checked Back:abcdefghijk

  8. Source 8: Authentication & access control LangChain · checked Back:abcdefg

  9. Source 9: A2A endpoint in Agent Server LangChain · checked Back:abcde

  10. Source 10: MCP endpoint in Agent Server LangChain · checked Back:abcd

  11. Source 11: No-code agents with LangSmith Fleet LangChain · checked Back:abcdefghi

  12. Source 12: How to Build a Multi-LLM AI Agent with Kong AI Gateway and LangGraph Kong · checked Back to text

  13. Source 13: Agents in Catalog - Kong Docs Kong · checked Back:abcd

  14. Source 14: Agent platform comparison LangChain · checked Back:abcde

  15. Source 15: LangSmith Deployment LangChain · checked Back:abcd

  16. Source 16: Remote MCP servers LangChain · checked Back:abc

  17. Source 17: AI Auth Strategies - Kong AI Gateway docs Kong · checked Back:abcde

  18. Source 18: Access & oversight LangChain · checked Back:abcdef

  19. Source 19: Kong Pricing & Plans Kong · checked Back:abcdefghijklm

  20. Source 20: Kong AI Gateway 2.0 Is Now GA - And It's Already Moving Faster Kong · checked Back:abc

  21. Source 21: LangGraph Platform is now Generally Available: Deploy & manage long-running, stateful Agents LangChain · checked Back:ab

  22. Source 22: Now GA: LangSmith Agent Builder LangChain · checked Back:ab

  23. Source 23: LangSmith Fleet changelog LangChain · checked Back:abc

  24. Source 24: Route A2A agent traffic through AI Gateway - Kong Docs Kong · checked Back:ab

  25. Source 25: Kong AI Gateway | Kong Docs Kong · checked Back:abcde

  26. Source 26: Essentials LangChain · checked Back:abcd

  27. Source 27: Route A2A traffic through AI Gateway - Kong Docs Kong · checked Back:abc

  28. Source 28: LangSmith control plane LangChain · checked Back:abc

  29. Source 29: Trace with OpenTelemetry LangChain · checked Back:ab

  30. Source 30: Cloud (SaaS) LangChain · checked Back:abc

  31. Source 31: Konnect SSO - Kong Docs Kong · checked Back to text

  32. Source 32: Authentication methods LangChain · checked Back to text

  33. Source 33: Kong AI Gateway Policies - Kong Docs Kong · checked Back:ab

  34. Source 34: Request Termination - Configuration Reference - Policy | Kong Docs Kong · checked Back to text

  35. Source 35: Overview (administration) LangChain · checked Back to text

  36. Source 36: Konnect and Dev Portal audit logs - Kong Docs Kong · checked Back:ab

  37. Source 37: Audit logs LangChain · checked Back:ab

  38. Source 38: FIPS 140-3 compliance in AI Gateway - Kong Docs Kong · checked Back:ab

  39. Source 39: Trust Center - Kong Inc. Kong · checked Back:ab

  40. Source 40: Trust Center - LangChain LangChain · checked Back:ab

  41. Source 41: Kong AI Gateway changelog - Kong Docs Kong · checked Back to text

  42. Source 42: Assistants LangChain · checked Back to text

  43. Source 43: Role-based access control LangChain · checked Back to text

  44. Source 44: AI PII Sanitizer - Policy | Kong Docs Kong · checked Back to text

  45. Source 45: Manage agent settings LangChain · checked Back to text

  46. Source 46: Manage workspace administration LangChain · checked Back to text

  47. Source 47: Enable additional LangSmith features LangChain · checked Back to text

  48. Source 48: MCP version support - Kong AI Gateway docs Kong · checked Back:ab

  49. Source 49: The Agent Gateway for Secure, Observable Agent-to-Agent Communication (Kong) Kong · checked Back:ab

  50. Source 50: Deploy other frameworks LangChain · checked Back to text

  51. Source 51: Geographic regions - Kong Docs Kong · checked Back to text

  52. Source 52: Dedicated Cloud Gateways - Kong Docs Kong · checked Back to text

  53. Source 53: Serverless Gateways - Kong Docs Kong · checked Back to text

  54. Source 54: Bring Your Own Cloud (BYOC) LangChain · checked Back:ab

  55. Source 55: Support Plans LangChain · checked Back:ab

  56. Source 56: Deploy your app to cloud LangChain · checked Back to text

  57. Source 57: AI Gateway providers - Kong Docs Kong · checked Back:ab

  58. Source 58: LangSmith Fleet LangChain · checked Back to text

  59. Source 59: LangSmith Observability LangChain · checked Back to text

  60. Source 60: Why Blocks? Blocks.ai · checked Back to text

  61. Source 61: What is Blocks? Blocks.ai · checked Back to text

  62. Source 62: Your company's private network Blocks.ai · checked Back to text

  63. Source 63: Network requirements Blocks.ai · checked Back to text

  64. Source 64: Solutions: Agent sprawl Blocks.ai · checked Back to text

  65. Source 65: Solutions: Partner networks Blocks.ai · checked Back to text

  66. Source 66: Pricing Blocks.ai · checked Back to text