Skip to content

LLM Projects collection type in Strapi

Summary

Introduce a dedicated LLM Projects collection type in Strapi to manage AI/LLM-based projects in a structured and scalable way.

This feature will allow administrators to create, categorize, and manage Large Language Model (LLM) projects with AI-specific metadata such as model details, duration, outcomes, future plans, and technical stack.

It improves content organization, scalability, and dynamic frontend rendering.

Problem to Solve

Currently, LLM projects are either:

  • Managed as generic content without AI-specific structure, or
  • Hardcoded in the frontend

This creates limitations in scalability, filtering, and content management.

User Story:

As a CMS administrator, I want a structured LLM Projects content type in Strapi so that I can manage AI-related projects dynamically without modifying frontend code.

Intended Users

  • CMS Administrators
  • Content Editors
  • AI/ML Project Managers
  • Frontend Developers
  • End users browsing LLM projects

User Experience Goal

Administrators should be able to:

  • Create and manage LLM projects in Strapi
  • Add AI-specific details (model used, outcomes, plans, metrics)
  • Categorize projects (e.g., NLP Research, Model Optimization)
  • Publish and control visibility
  • Update project content without code changes

Frontend should dynamically render projects via Strapi API.

Proposal

1️⃣ Create Collection Type: LLM Project

Suggested Fields:

  • title (Text)
  • slug (UID)
  • shortDescription (Text)
  • longDescription (Rich Text)
  • startQuarter (Enum: Q1, Q2, Q3, Q4)
  • startYear (Integer)
  • endQuarter (Enum)
  • endYear (Integer)
  • status (Enum: Planned, Ongoing, Completed)
  • modelUsed (Text)
  • techStack (Repeatable text or component)
  • coverImage (Media)
  • featured (Boolean)

2️⃣ Create Supporting Components

Key Outcome Component

  • title (Text)

Future Plan Component

  • title (Text)

Attach as repeatable components to LLM Project.

3️⃣ Create Category Collection (Optional but Recommended)

Collection Type: LLM Project Category

Fields:

  • title
  • slug
  • description
  • Relation: One-to-Many with LLM Project

This enables grouping like:

  • Telugu Language Foundation
  • Advanced NLP Research
  • Model Optimization

4️⃣ Expected Admin Flow

  1. Go to Content Manager
  2. Create LLM Project Category
  3. Create LLM Project
  4. Attach outcomes & future plans
  5. Assign category
  6. Save & Publish

5️⃣ Expected API Usage

Fetch categories with projects:

/api/llm-project-categories?populate[projects][populate]=*

Or fetch all projects:

/api/llm-projects?populate=*

Further Details

Benefits

  • Structured AI project management
  • Eliminates hardcoded frontend content
  • Enables filtering by category or status
  • Future-proof architecture
  • Reusable components

Trade-offs

  • Increased schema complexity
  • Requires proper population in API queries

Security or Permissions Considerations

  • Enable find and findOne for public role (if public data)
  • Sensitive API keys must NOT be stored in project entries
  • Draft/Publish workflow should be enabled
  • Optional: Role-based editing restrictions

No elevated permissions required beyond CMS editing access.

Documentation Impact

Update:

  • CMS data model documentation
  • API usage examples
  • Frontend integration guide
  • Content entry guidelines

Add section:

Managing LLM Projects in Strapi

Testing Considerations

Unit Tests

  • Field validation
  • Enum validation
  • Required field checks

Integration Tests

  • API population works correctly
  • Category relations return expected data

Edge Cases

  • Project without outcomes
  • Empty tech stack
  • Large descriptions
  • Unpublished entries

End-to-End Tests

  • Create → Publish → Fetch → Render

Milestone & Labels

Milestone:

  • AI CMS Enhancement Phase

Suggested Labels:

  • enhancement
  • cms
  • strapi
  • feature-request
  • ai
  • llm
  • backend

Related Issues or References

  • Frontend LLM project display feature
  • CMS restructuring discussions
  • AI portfolio roadmap