Contact Us
Software Units Layer
In Reqode, you can link specifications and source code files. By linking specs to code, Reqode will be able to provide relevant, focused context for both embedded AI assistants and external AI agents through the MCP server.
Tracing specs to source code unlocks capabilities that go beyond traditional requirements management systems:
- See every commit that touches a linked artifact.
- Automatically update the test status of requirements when related code changes.
- AI assistant analyzes commits and detects inconsistencies with the requirements.
- Updating requirements, specifications and tests using AI-assistants based on the current state of the source code.
- AI-assisted code generation based on requirements and specifications.
- The MCP server provides external AI agents (e.g., IDE co-pilots) with up-to-date documentation and requirements based on file paths.
- An IDE plugin that provides developers with current tasks and documentation for each file and related logic.
How It Works
Source code information is collected through integration with your version control system. If your software consists of multiple subsystems, you can customize the integration for each subsystem and its repository.
Once the integration is configured, Reqode will retrieve commit information and source code from the selected branch. If you don't plan to use AI for code analysis, you can disable source code download and keep only commits and file paths for artifacts linking.
The link between requirements/specifications and source code files is created using an intermediate artifact called a Unit.
A unit is an architectural object that groups together source code files that perform a single functionality in the system. Units allow you to manage sets of files as one element of the software architecture. E.g., in some UI frameworks, components always consist of several files. From the requirements perspective, the internal structure of these files is not important, so all files for a component are grouped into a single Unit representing that component.
In Reqode, you can configure different types of Units and specify instructions for the AI assistant for analysis and code generation for each type.
Units as Blueprints for AI Coding
As an intermediate artifact between a feature and the source code, a unit defines architectural decisions that an AI agent must follow during both initial implementation and future changes.
Code traceability through software units layer mitigates hallucinations and ensures stricter control over code generation, thereby minimizing unwarranted deviations introduced by the AI coding agents.
From Requirements to Code and Back
This approach allows Reqode to be used not only for requirements management, but also for verification and validation of requirements implementation, as well as for the implementation itself.