Akita Software's platform specializes in building application programming interface (API) models from existing API traffic, allowing users to compare specifications with observed behavior from test, staging, or production environments. The company's platform builds dynamic models of API behavior in order to automatically generate specifications for any API, update its specifications, and discover and document endpoints. The purpose of the platform is to enable businesses to deploy a service observability solution without interfering with systems, as well as improve technical issue diagnosis and service delivery processes.
In designing its solution, the company took inspiration from program synthesis approaches that learn programs by observing interactions with a system, such as those adopted by DemoMatch. Akita has also taken inspiration from invariant generation approaches such as those adopted by Daikon, whose platform automatically learns properties about programs through trial and error testing of properties. The company aims to innovate on these techniques by analyzing traces of network calls instead of program traces in order to enable candidate-based generation of invariants about API behavior.
With Akita, users can gain an up-to-date understanding of endpoints, data, and cross-service dependencies throughout their system without having to monitor services individually. The platform is operated by deploying an Akita agent onto a network or into Kubernetes clusters. Akita can monitor any service in CI/CD (continuous integration/continuous delivery), staging, or production and summarize traffic to allow users to explore per-endpoint behavior. Akita's features include the following:
- Service graph visualization: visualized metrics showing performance, counts, data types, data formats, and response codes
- Identification of data formats and sensitive data
- Creation of API traffic logs through automatic monitoring of API endpoint behavior
- Automated regression detection: automated regression detection comprises users' ability to identify potential breaking and security-critical changes. This feature combines automatic alerting, detection through structured modeling, and issue detection capabilities at the pre-production stage
- Generation of OpenAPI specifications
Akita Software's platform is eBPF-based, meaning it is a "black box" system. As a result, installing Akita does not require the implementation of changes to code or proxying, enabling it to be deployed across all parts of a system, including legacy APIs.
The platform focuses on high-level, developer-centric reporting; according to the company, it can be particularly useful to developers with large workloads who may be building on a pre-existing code base without specialized knowledge of its structure. Users are not required to have any observability expertise to operate the Akita dashboard.
Users can run a limitless amount of traffic via Akita and get an explorable view of their API endpoints, which can be filtered according to data type and other parameters. Users can then export the result as a complete OpenAPI 3 specification that is compatible with other API tools.
Akita's pricing plans are based on the amount of traffic sent by the user and how long they wish to preserve the data. During the platform's beta stage, the company offered Private Beta, Free, Pro, Teams, and Enterprise subscription plans. The Private Beta and Free plans came at no cost, while the Pro and Teams plans were charged at $150 and $300, respectively. The Enterprise plan for customers with the largest traffic demands was offered as a custom pricing option, depending on the scale required, and came with the most complete set of the platform's features.