# Zine Case Study

How [Zine](https://www.zine.ai) demonstrates what's possible when you build on Graphlit.

{% hint style="success" %}
**Zine is proof Graphlit is production-ready.** Built 100% on Graphlit, Zine shows how to create team memory that actually works - connecting 20+ tools and making everything searchable with AI.
{% endhint %}

***

## What is Zine?

**The product:** Your team's memory, everywhere you work.

**What it does:**

* Connects Slack, Gmail, GitHub, Notion, Linear, Jira, Google Drive, and 20+ other tools
* Makes everything searchable with semantic AI (not just keyword matching)
* Provides MCP server so your IDE can access your team's knowledge
* Automatic sync across all connected sources (configurable polling)

**The tech stack:** Next.js frontend + Graphlit for everything else.

{% hint style="info" %}
**Philosophy:** Focus on user experience, not infrastructure. Graphlit handles ingestion, processing, search, and AI. Zine builds the product.
{% endhint %}

***

## What Graphlit Enabled

Building on Graphlit let Zine focus on user experience, not infrastructure.

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>What Zine built</strong></td><td>User experience<br>Next.js frontend<br>Product features<br>Deployment</td></tr><tr><td><strong>What Graphlit provides</strong></td><td>30+ feeds<br>OCR &#x26; transcription<br>Embeddings &#x26; search<br>Entity extraction<br>Knowledge graph<br>Multi-tenant isolation<br>Webhook management<br>Automatic sync</td></tr></tbody></table>

{% hint style="success" %}
**Result:** Ship features, not infrastructure. Zine went from idea to production in weeks, not months.
{% endhint %}

***

## The Development Story: AI-Assisted Velocity

Zine demonstrates what's possible when you combine Graphlit's platform with modern AI coding tools.

**Timeline:**

* **March:** First line of code
* **6 weeks later:** Working MVP
* **June (3 months):** Paying customers
* **Today:** Production SaaS with multiple workspaces

**How it was built:**

* **One developer** (not even a frontend specialist)
* AI coding tools: Claude Code, Cline, Factory Droid
* Graphlit TypeScript SDK
* "Vibe coding" approach - describe what you want, AI implements it

{% hint style="info" %}
**The 2025 way to build:** One developer + AI tools + great SDKs = what used to require a full team. You don't need to be a full-stack expert. The AI handles implementation, Graphlit handles infrastructure.
{% endhint %}

**What this enabled:**

* Focus on product experience, not infrastructure plumbing
* Iterate rapidly based on user feedback
* Ship features in days, not weeks
* Solo developer building what used to require a full engineering org

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>Traditional approach</strong></td><td>6-12 months to build infrastructure<br>Large engineering team<br>Stitching together 10+ services<br>Custom OAuth for each connector<br>Vector DB management<br>Scaling challenges</td></tr><tr><td><strong>Graphlit + AI coding</strong></td><td>6 weeks to MVP<br>One developer (AI-assisted)<br>One SDK, one platform<br>30+ feeds work immediately<br>Production-ready from day one<br>Focus on user experience</td></tr></tbody></table>

{% hint style="success" %}
**Key insight:** The right infrastructure abstractions unlock AI-assisted development. Graphlit's SDK is designed to be AI-coding-friendly.
{% endhint %}

***

## The Problem Zine Solves

### Before Zine

**Knowledge scattered everywhere:**

* 💬 Slack conversations (decisions, discussions)
* 📧 Gmail threads (customer communications)
* 📅 Meeting recordings (context, action items)
* 📝 Notion docs (specs, wikis)
* 🐛 Linear/Jira issues (product roadmap)
* 💻 GitHub discussions (technical decisions)

**The pain:**

* "What did Sarah from Acme Corp say about pricing?" → Check 5 tools manually
* "Why did we make this decision?" → Lost in Slack history
* "What were the action items from last week's call?" → Scroll through recordings

**Time wasted**: Hours per week per person searching for context.

***

### After Zine

**One search across everything:**

```
Search: "Acme Corp pricing concerns"

Results:
📧 Email (Oct 12): Sarah initial inquiry  
💬 Slack (Oct 13): Team discussion  
📅 Meeting (Oct 15): Sales call transcript  
📝 Notion (Oct 16): Deal notes  
🐛 Linear (Oct 17): Pricing feature request
```

**Time saved**: Seconds to find anything.

**Powered by Graphlit's semantic memory.**

***

## How Zine Works

### Multi-Source Ingestion

**Zine creates Graphlit feeds** for each connected tool (Slack, Gmail, GitHub, etc.).

**Example:** Connect Slack channel → Graphlit handles:

* OAuth token refresh
* Message polling
* Thread preservation
* File extraction
* Deduplication

**Result:** Connect once, syncs forever. No webhook infrastructure needed.

***

### Entity Extraction (Knowledge Graph)

**Zine enables entity extraction in workflows.**

**Graphlit automatically extracts:**

* People (e.g., "Sarah Chen, CTO")
* Organizations (e.g., "Acme Corp")
* Places (e.g., "San Francisco")
* Events (e.g., "Q4 Planning Meeting")
* Products mentioned

**Use case:** "Show me all interactions with Acme Corp" → Graphlit finds mentions across Slack, email, docs, meetings.

***

### Multi-Tenant Architecture

**Pattern:** Each Zine workspace = separate Graphlit environment.

**Benefits:**

* Complete data isolation per customer
* Independent scaling
* Usage tracking per workspace
* Easy data deletion

***

### Unified Search

**Zine uses Graphlit's hybrid search** (vector + keyword) across all connected sources.

**Filters available:**

* By source (Slack, Gmail, docs, etc.)
* By date range
* By entities (people, organizations)
* By content type

**Graphlit handles:**

* Vector embeddings
* Search ranking
* Result deduplication
* Cross-source search

***

### AI Chat with Context

**Zine creates Graphlit conversations** that automatically retrieve relevant context from the user's knowledge base.

**Features:**

* Streaming responses
* Automatic context injection
* Multi-turn memory
* Source citations

***

### Meeting Intelligence

**Zine ingests meeting recordings** via Graphlit audio processing.

**Graphlit provides:**

* Audio transcription (Deepgram, AssemblyAI)
* Entity extraction from transcripts
* Searchable meeting content
* Summary generation

***

## What Builders Can Learn from Zine

### 1. AI Coding + Great SDKs = Velocity

**The MVP was built in 6 weeks with AI coding tools** (Claude Code, Cline, Factory Droid). Paying customers in 3 months.

**Why it worked:**

* Graphlit SDK is AI-coding-friendly (clear patterns, TypeScript types)
* Platform handles all the hard infrastructure problems
* Focus 100% on user experience and product

### 2. Multi-Tenant Architecture

Use separate Graphlit environments per customer for complete data isolation.

### 3. Focus on Your Unique Value

Don't build OAuth management, feed polling, vector search, or transcription. Let Graphlit handle infrastructure.

**Time saved:** Months of infrastructure work → weeks building product features.

### 4. Entity Extraction from Day One

Enable entity extraction in workflows - it unlocks powerful queries like "Show me all interactions with Acme Corp."

### 5. Production Patterns

* Retry logic for transient failures
* Usage tracking per customer
* Background jobs for long operations
* Client caching for performance

{% hint style="success" %}
**Zine's lesson:** Build with platforms, not from scratch. Use AI coding to accelerate. Ship fast, iterate fast.
{% endhint %}

***

## Zine's Tech Stack

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>Frontend</strong></td><td>Next.js<br>TypeScript<br>Tailwind CSS</td></tr><tr><td><strong>Backend</strong></td><td>Next.js API routes<br>Graphlit TypeScript SDK<br>Clerk (auth &#x26; user metadata)<br>Redis (caching)</td></tr><tr><td><strong>Infrastructure</strong></td><td>Vercel (hosting)<br>Graphlit (content/AI)</td></tr></tbody></table>

{% hint style="success" %}
**What Zine didn't build:** Vector databases, embedding pipelines, search infrastructure, 30+ feed connectors, OAuth management, audio transcription, entity extraction, knowledge graphs.

**Result:** Focus 100% on user experience and product features.
{% endhint %}

***

## Try Zine or Build Your Own

**Want to see it in action?**

* [Try Zine](https://www.zine.ai) - Free to start

**Want to build your own in weeks, not months?**

* [Get Started with Graphlit](/getting-started/quickstart.md) - Build your first agent in 7 minutes
* [Use Case Library](/api-guides/use-cases.md) - 117+ code examples

{% hint style="info" %}
**AI-coding-friendly SDK:** Use Claude Code, Cline, Cursor, or Factory Droid with Graphlit's TypeScript, Python, or .NET SDKs. Clear patterns, comprehensive types, great docs.
{% endhint %}

***

**Zine went from idea to paying customers in 3 months. What will you build?**


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.graphlit.dev/examples/zine-case-study.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.