Content & AIv2 in development

Khottba v2

A platform that links generated content to trusted sources

Context

It is not enough for AI-generated content to sound correct, especially in sensitive topics; every claim needs a clear source. Version one proved the idea, and version two rebuilds it in a clearer, easier-to-maintain way.

Problem

AI-generated content in sensitive domains needs a guarantee that every statement traces back to a trusted source.

Solution

A platform that links every passage to its source, supports multiple AI providers, exports content to Word, and uses tests to protect its main functions.

Highlights

  • The source of every passage is stored in the system
  • Version one functions protected during the rebuild
  • System parts separated to make testing and development easier

Implementation

  1. 1

    Protect the previous version's functions

    Tests recorded how version one worked before rebuilding, so no important function would be lost.

  2. 2

    Split the system into clear parts

    Business rules, the database, AI services, and export are separated so each part can be tested and improved on its own.

  3. 3

    Link content to its source

    The relationship between each passage and its trusted source is stored in the data instead of adding references manually later.

  4. 4

    Prepare content for real use

    Content can be exported to Word, and version one data can be moved safely through a documented process.

Stack

Next.js 16MonorepoPrismaAI ProvidersVitest

Results & numbers

5
independent parts that make the system easier to test and improve
3
test levels covering functions, system parts, and user journeys
100%
of content linked to its source inside the system