Tyk API Management

What do you like best about the product?

We major on Kubernetes, and Tyk makes it straightforward to manage that infrastructure as code while covering both AI and APIs. Our team lives in CI/CD pipelines, not GUIs, and Tyk really fits that way of working. As we roll out more LLM capabilities internally, we expect to make even more use of the Tyk AI Gateway.

What do you dislike about the product?

They really need better documentation. The information is there, but it’s confusing and not laid out in a clear way. I also wish the operator were more fully featured, especially when it comes to the Developer Portal.

What problems is the product solving and how is that benefiting you?

We needed a fast way to get a single cohesive view of API and AI consumption and we needed a single data plane to enforce access. Now all of that is visible in Open Telemetry dashboards, so we're getting on top of governance and reducing risks. We can also see more about who is spending what on AI

What do you like best about the product?

The user interface makes it easy to configure the API and manage analytics. It also offers multiple middleware options for request and response transformations. The configuration is very granular, yet straightforward to integrate, and it delivers efficient performance. The user support team is also quick to respond.

What do you dislike about the product?

Enterprise features such as request encryption and signing have not yet been integrated. It would also be better if observability-related features were added directly in the dashboard itself, so users can manage all API configuration and see their running data analytics and logs within the same dashboard.

What problems is the product solving and how is that benefiting you?

Tyk performs quickly and offers a variety of options for managing APIs, including the UI, API, and operator. The dashboard also provides a good user experience for consumers.

What do you like best about the product?

What I like best about Tyk is its flexibility, lightweight architecture, and ability to support scalable API management across hybrid and cloud environments. Its API-first approach, strong security capabilities, flexible policy management, and support for automation make it well suited for enterprise platforms. I also value its Kubernetes compatibility, deployment flexibility, and integration with CI/CD pipelines, which help teams standardise API onboarding, improve operational efficiency, and enable self-service adoption across multiple teams.

What do you dislike about the product?

While Tyk offers strong flexibility and API management capabilities, some areas could be improved, particularly the usability of the management interface, troubleshooting experience, and visibility into operational issues across distributed deployments. More comprehensive documentation and practical examples for complex enterprise use cases, along with simpler configuration and debugging workflows, would help reduce the learning curve and improve the overall developer and operator experience.

What problems is the product solving and how is that benefiting you?

Tyk solves the challenge of managing, securing, and governing APIs consistently across distributed teams and hybrid cloud environments. It provides a centralised approach to API authentication, authorisation, traffic management, rate limiting, and access control, while enabling teams to expose APIs in a standardised and controlled manner. This benefits me by improving platform security, simplifying API onboarding, promoting self-service adoption, reducing manual operational effort through automation, and enabling consistent governance across teams. It also supports scalability and operational efficiency while allowing flexibility in deployment and integration with existing enterprise infrastructure and CI/CD pipelines.

What do you like best about the product?

Tyk has been an integral part of our API management and transformation journey for a couple of years now. Tyk’s biggest strengths are its performance, observability, and evolving capabilities for AI integration.

We use Tyk as a self-managed platform, replacing MuleSoft as our internal API gateway. So far 16-20 teams have adopted the gateway, with more in the queue.

Performance: Tyk is an ultra-fast, low-latency API gateway. In our testing, the platform has been validated at 10,000 requests per second, making performance a key strength for enterprise API traffic.

Logging and Monitoring: The Tyk Dashboard provides valuable visibility into API traffic, including success and error rates at both API and API-key levels. This helps teams monitor API health, identify issues, and understand usage patterns more effectively.

AI and MCP Capabilities: Tyk’s introduction of MCP support is particularly promising. The ability to extend API gateway security and governance capabilities to AI tools and agent-based interactions, while maintaining authentication, authorization, logging, and monitoring, is a valuable step towards secure enterprise AI adoption and Intelligence at scale.

Overall, Tyk combines high performance and operational visibility with capabilities that can support the next generation of API and AI-driven platforms.

We are also looking to use Tyk’s MCP capability in our AI and Intelligence initiatives, to provide secure communication from agentic chatbots to the different tool sets available across the organization.

What do you dislike about the product?

Tyk works well as a federated model, but the data plane components need to be tested carefully to suit your specific performance needs. With that effort up front, the results have been very strong for us.

What problems is the product solving and how is that benefiting you?

We run a strong microservices ecosystem across the organization, and operating at that scale requires a gateway with best-in-class authentication and performance.

Our previous integration toolset had two main issues. First, it created heavy vendor lock-in. Second, it relied on a lot of custom tooling built around specific needs, which deepened that dependency and made the platform harder to evolve. We wanted to move away from this model. We also wanted a Kubernetes-native gateway that runs in our Kubernetes environments out of the box.

Tyk met these needs: it is Kubernetes-ready, its performance benchmarks are well tested, and it gave us a clear path away from vendor lock-in.

We built reusable pipelines for API onboarding and for infrastructure setup, working consistently across both cloud and on-premises environments. This cut API onboarding time by around 70%. We also reduced our dependency on any single tool, which brought down the cost of running the platform by around 40%.

Quantified benefits

API onboarding time reduced by around 70%
Platform cost reduced by around 40%
Consistent deployment across cloud and on-premises
Reduced vendor lock-in and less custom tooling to maintain

What do you like best about the product?

We were looking for a platform that handles AI and API workloads. We moved about 140 internal and partner APIs onto the Tyk Gateway over six months. What I like most is how much we can do declaratively — rate limits, quotas and JWT auth are set per API in config we keep in Git, so changes go through the same review as our code. Latency overhead has been negligible for our traffic (we measured low single-digit milliseconds at p95). The Developer Portal also cut partner onboarding from a ticket-and-wait process to self-service.

What do you dislike about the product?

The learning curve on advanced policy and plugin configuration was steeper than we expected — the docs cover the common cases well, but we needed support's help for a couple of edge cases around custom Go plugins. Planning major-version upgrades also takes care across multiple environments. Support was responsive when we raised both, but better worked examples for complex set-ups would help.

What problems is the product solving and how is that benefiting you?

We’re building cutting edge AI infrastructure for a global financial institution. Before Tyk, each team secured and exposed its own APIs, so auth, logging and rate limiting were inconsistent and audits were painful. Now every API goes through one gateway with standard policies, which gives our security team a single place to check. Partner integrations that took three to four weeks to set up now take a few days, and we've retired home-grown proxies we were maintaining.