1: Build a Chat Agent Platform for Developers
Your challenge is to build a plug-and-play platform that helps developers create custom chat
agents in just a few steps. These agents should be powered by LLMs (like Groq) and work with
content stored in Contentstack.
What You Need to Build
1. LLM Model API - An API that:
● Sends and receives chat messages (supports streaming responses)
● Lets developers switch between different LLM providers (like OpenAI, Anthropic, etc.)
● Connects to Contentstack’s Delivery APIs (via MCP) to fetch content based on the
user’s question
2. Chat SDK - A lightweight SDK that:
● Makes it easy to embed chat agents in any website or app
● Works well with React (ideally using React Hooks)
● Connects to your LLM Model API to power conversations
Developers should be able to use your platform to build domain-specific chat agents without
needing to understand all the backend logic.
For example, on a travel website using Contentstack, the chatbot should respond to: “What
tours are available for Italy?”
Your platform should take care of:
● Understanding the question and querying the right content
● Sending back a useful response — all with minimal setup
Contentstack Products Purpose
CMS To model and manage structured content
Launch To build or embed your chat UI using its APIs
Contentstack MCP To connect to the content delivery layer and pull live content
Developer Hub (OAuth) To register the app and access stack-level APIs
2: Build a Smart, Embeddings-Based Search App for Contentstack
Your task is to build a Contentstack Marketplace App that adds semantic search to any
stack. Unlike basic keyword search, this app should understand the meaning behind what users
are looking for and return the most relevant content — even if the exact words don’t match.
What You Need to Build
1. Semantic Search Engine
● Accept free-form, natural language search queries (e.g., “red high top sneakers with
stripes”)
● Return the most relevant entries using semantic similarity, not just keyword matching
● Rank results based on how closely they match the intent of the query
2. Embeddings Integration
● Use LLM-based embeddings (like OpenAI, Cohere, or open-source models)
● Represent content entries and queries as vectors in a shared space
● Use vector similarity to compare and rank entries
3. CMS Sync & Indexing
● Keep your search index up to date with content changes
● Use Contentstack Webhooks to automatically update the index when entries are
created, updated, or deleted
● Ensure low-latency updates so new content is searchable quickly
4. Marketplace App UI
Build a UI inside Contentstack (hosted via Launch) where users can:
● Enter search queries and see ranked results with similarity scores
● Optionally filter by content type or locale
5. Scalability
● Your search system should handle thousands of entries across multiple content types
● Use tools like FAISS, Weaviate, or Pinecone for fast and efficient vector search
Contentstack Products Purpose
CMS To store and manage all your content entries
Launch To host your Marketplace App and UI inside Contentstack
Automate To keep the search index synced with content changes
3: Smart Find and Replace for Editors
Build a Contentstack Marketplace App that helps editors perform smart, context-aware find
and replace across their content — going beyond basic text replacement.
What You Need to Build
1. Contextual Replacement
● Don’t just replace words — understand the meaning.
● Example: Replace “Gemini 2.5 Pro” with “Claude Sonnet”, not “Claude 2.5 Pro”.
2. Smart Link Updates
● When replacing text inside hyperlinks, update both the display text and the URL.
● Example: Change “Google” linking to https://google.com → “Bing” linking to
https://bing.com.
3. Named Entity Replacement
● Find and replace named entities like company names, emails, etc.
● Example: Replace “Alpha Company” and mailto:contact@alphacompany.com with
“Omega Corporation” and new email.
4. Deep Content Coverage
Your app must scan and edit Rich Text fields, Tables and nested components, Links, metadata,
and custom fields
5. Brandkit Integration
Use Brandkit to enforce brand-approved terms. Avoid banned or outdated phrases. Maintain
consistent tone and style
Contentstack Products Purpose
CMS Access and edit content entries
Brandkit Suggest smart, brand-aware replacements
Launch Host your app UI within the Contentstack dashboard
Automate Trigger follow-up actions
4. Build a Locale Variants App for Translations
Your challenge is to build a Marketplace app that helps teams manage multilingual content
using variant groups and fallback logic — all within Contentstack.
What You Need to Build
1. Variant Manager UI - An interface where users can:
a. Create variant groups (e.g., Languages)
b. Add locales (e.g., Hindi, Marathi) with fallback rules (e.g., Marathi → Hindi →
English)
2. Entry Sidebar Extension - A sidebar that:
a. Shows all variants for a given entry
b. Highlights missing translations
c. Lets users generate or override translations (manually or via AI)
d. Supports bulk variant creation for content types
3. Personalize API Integration - Use the Variants API (via Contentstack Personalize) to:
a. Store and retrieve locale-specific versions
b. Attach variant params like marathi_hindi_en to each locale
c. Implement fallback logic when a field is missing
Developers and content teams should be able to:
● Define a fallback hierarchy for locales
● Deliver the closest available translation without manual checks
● Use the variant param in frontend code to fetch the correct version
Contentstack Products Purpose
CMS To manage localized entries
Launch To visualize and preview variant-driven pages
Personalize API To store and resolve variant fallback logic
Developer Hub (OAuth) To register the app and access stack-level APIs
