Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Kong announced Kong Context Management as generally available in Kong AI Management on September 30, 2026. It is designed to let AI agents discover and use enterprise APIs, MCP servers and other data sources without loading every possible tool definition into each model request.
The approach combines a compact MCP interface called Code Mode with Context Mesh, a mediation layer for discovering APIs, generating agent tools, executing operations and enforcing gateway policies.
Why large MCP tool catalogs create a production problem
Model Context Protocol (MCP) can expose an API by representing each operation as a separate tool. That is straightforward for a small service, but a large enterprise API portfolio can produce hundreds of tool definitions. Sending that entire catalog to a model on every request consumes context-window capacity before the agent has done useful work.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
- Higher token use: Tool schemas occupy prompt space and must be processed repeatedly.
- Harder selection: The model has to distinguish among many similar operations, increasing the chance of choosing the wrong one.
- More latency: Larger prompts take longer to transmit and evaluate.
- More integration work: Each exposed operation needs consistent authentication, authorization and policy handling.
Kong’s central proposition is that broad enterprise access should be discoverable at runtime rather than placed wholesale in every prompt. The model receives enough capability to search and execute, while detailed schemas are retrieved only when they are relevant.
How Kong Code Mode reduces MCP overload
Code Mode presents agents with a compact MCP surface instead of a separate tool definition for every API operation. The agent can search the available API surface, retrieve the schema for a matching operation and execute it when needed.
- Search: The agent queries the available enterprise API capabilities through the compact interface.
- Retrieve: It obtains the schema for the operation or operations that match the task.
- Execute: It invokes the selected API operation through Kong’s managed path.
- Compose: For a multi-step task, it can combine calls in code using loops, conditionals and error handling.
Composition is important because a conventional agent may need a separate model-to-tool exchange for every step. Code Mode can express a sequence of operations together, reducing repeated reasoning and round trips. Kong presents this as a design intended to improve efficiency, but the announcement does not provide independent benchmark results, customer latency measurements or a quantified token reduction.
What Context Mesh adds
Kong’s Context Mesh launch on February 10, 2026 describes a broader control and discovery layer around agent access. Kong documentation calls it “the mediation layer between agents and APIs.” Its stated functions include:
| Capability | What Kong describes |
|---|---|
| API discovery | Automatically discover APIs managed in Kong Konnect. |
| Curated context | Assemble toolkits that select the context an agent should receive for a task or role. |
| MCP generation | Generate MCP tooling from managed APIs without building each tool integration by hand. |
| Deployment | Deploy the resulting agent access through Kong AI Gateway. |
| Access inheritance | Carry existing access controls into the agent-facing layer. |
| Runtime policy | Apply policy enforcement while an agent discovers and calls capabilities. |
| Future context sources | Kong’s direction extends beyond REST APIs toward event streams and other enterprise sources. |
This makes Context Mesh different from an MCP adapter that only translates API endpoints into tools. The intended system decides what an agent can discover, how that capability is packaged, where calls are executed and which controls apply at runtime.
How the pieces fit together
Context Management is the umbrella capability in Kong AI Management. Code Mode addresses the model-facing problem: expose a small interface and fetch details on demand. Context Mesh addresses the enterprise-facing problem: discover and organize capabilities, generate MCP access, and mediate calls. Kong AI Gateway supplies the deployment and governance path for those interactions.
The combination is aimed at production agents that need access to many back-end systems but should not receive unrestricted context. An organization could curate different toolkits for different agents, inherit API permissions, and enforce policies as requests pass through the gateway. The announcement does not specify a single required architecture for every data source or agent framework.
Governance matters as much as interoperability
MCP can make systems interoperable, but interoperability alone does not answer which agent may access which operation, what credentials it may use, or how calls should be monitored. Kong positions Context Management as an infrastructure layer that combines:
- Runtime discovery instead of static exposure of every operation.
- Selective context and toolkit curation.
- Authentication and inherited authorization controls.
- Policy enforcement at the gateway.
- A managed execution path for composed, multi-operation tasks.
Those controls are especially relevant when an agent can search a large API estate. Discovery must be constrained by identity and policy; otherwise, reducing prompt size could still leave the agent with access to capabilities it should not use.
Does it reduce token cost and latency?
Architecturally, Code Mode can reduce the amount of schema text placed in a prompt and can avoid repeated model-to-tool exchanges when several operations are composed in code. Those mechanisms can lower context processing and network overhead in suitable workloads.
However, Kong’s announcement and product materials do not publish a token-cost percentage, latency benchmark, independent test, customer performance measurement or public pricing. Results will depend on API size, schema complexity, search quality, model behavior, network distance, policy checks and how many operations a task requires. Treat efficiency as the product’s design goal, not as a guaranteed numerical outcome.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the announcement means for enterprise teams
Teams with large API portfolios
Organizations that already manage APIs in Kong Konnect may be able to reuse that inventory for agent access rather than creating a separate tool catalog. The practical evaluation question is whether discovery and curation accurately surface the right operations for each agent.
Security and platform teams
Inherited controls and runtime policy provide a place to apply identity, authorization and gateway rules to agent traffic. Teams should verify how existing policies map to generated MCP tools and how denied or partially completed multi-step calls are recorded.
Application developers
Code Mode’s support for loops, conditionals and error handling can simplify workflows that would otherwise require the model to orchestrate every API call. Developers still need to define safe boundaries, validate inputs and design recovery behavior for failed operations.
Procurement and architecture teams
Kong has not published pricing in the supplied announcement materials. A buying decision should therefore request current commercial terms, deployment requirements, observability details, supported data sources and service-level commitments directly from Kong.
Why Kong says the timing matters
Kong’s February 10, 2026 press release attributes a warning to a January 2026 Gartner report: 40% of agentic AI initiatives could be at risk of cancellation by 2027 if organizations fail to bridge the integration gap. That is a Gartner figure as reported by Kong, not an independently verified Gartner publication here.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Marco Palladino, Kong’s CTO and co-founder, described the rationale this way: “Organizations have spent years building APIs as the nervous system of the enterprise. Context Mesh allows us to reuse that investment to power agents instead of starting from scratch.”
Bottom line for adopters
Kong Context Management is generally available in Kong AI Management and targets a specific production bottleneck: exposing extensive enterprise capability without overwhelming the model with static MCP tools. Code Mode provides runtime search, schema retrieval and composable execution; Context Mesh supplies discovery, curated context, MCP generation and governance through Kong AI Gateway. The product’s efficiency claims are plausible from its architecture, but buyers should demand workload-specific measurements and pricing before treating lower token use or latency as established results.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
