Project Creation

Starting a new build or resuming a legacy project.

1. Creating a New Application

This workflow is for greenfield projects. You provide the intent; Klyve provides the architectural elaboration.

Step 1: Initialize

  • Navigate: File > New Project and select “Create from a New Specification”.
  • Jira Integration (Optional):
    • Go to Project Settings > Integrations to enable two-way backlog sync.
    • Input your URL, Username, and Jira API Token to authenticate.

Step 2: The Project Brief

You do not need a perfect technical document to start.

  • Text Mode: Paste a raw description (e.g., “A Java-based retail inventory system…“).
  • File Mode: Upload existing PDF or Word requirements documents. The better the spec the better the final deliverables.
  • Analysis: Klyve scans the input to identify core features and missing requirements.

Step 3: Lifecycle Assessment

Before generating specs, Klyve assesses the project type:

  • UX Decision: It asks if a UX Specification is needed. For backend-only component development, it will suggest skipping this.
  • Fast-Track: For very small or clearly defined scripts, it may recommend skipping the elaboration phase entirely and moving straight to development.

Step 4: Specification Ratification

Klyve generates the foundational documents for your review.

  • UX/UI Spec: User personas, journeys, and screen definitions.
  • Application Spec: Functional requirements and scope.
  • Technical Spec: Stack selection, database schema, API architecture.
  • Action: Review the drafts. Use “Refine” to request changes, then click “Approve” to lock the scope, generate coding standards, receive documented guidance for preparing the dev and test environments, generate build scripts and generate the project backlog.

2. Onboarding an Existing Codebase

This workflow is for legacy maintenance. Klyve ingests your local source code to build a context map.

Step 1: Back up your Codebase

  • AI may make mistakes, and this presents a risk that it may silently introduce new errors into your existing codebase or damage it. To mitigate this risk, always make sure you have a backup of your current codebase available before you make a change to it using Klyve.

Step 2: Select Target

  • Navigate: File > New Project and select “Work with an Existing Codebase”.
  • Input: Browse to the root folder of the application in which the user has made the code base files available.

Step 3: Pre-Flight Scan

Klyve performs a non-destructive read-only scan.

  • Safety: Verifies valid source code and checks for Git repositories.
  • Indexing: Builds a “Project Memory,” allowing the AI to understand the architecture without re-reading every file for every task.

Step 4: Reverse-Engineering

Klyve establishes a baseline by generating “Reverse-Engineered” specifications:

  • Baseline Docs: Creates Application and Technical Specifications that reflect the current state of the system.
  • Outcome: These documents become the source of truth for future changes.

Step 5: Maintenance Dashboard

Once indexed, you are taken to the Dashboard.

  • Status: Displays detected languages and codebase health.
  • Next Step: Define your coding standards, then navigate to the Project Backlog to begin adding bug reports or changes.
  • After a change: After the successful completion of each fix or change, remember to check in or back up your code again as a risk mitigation against potential AI mistakes in the next change.