Skip to content

Commit 56dc489

Browse files
committed
add Hypermode Graph MCP doc
1 parent c1946d9 commit 56dc489

File tree

2 files changed

+352
-1
lines changed

2 files changed

+352
-1
lines changed

docs.json

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -59,7 +59,8 @@
5959
"graphs/connect",
6060
"graphs/manage-schema",
6161
"graphs/manage-data",
62-
"graphs/manage-graph"
62+
"graphs/manage-graph",
63+
"graphs/graph-mcp"
6364
]
6465
},
6566
{

graphs/graph-mcp.mdx

Lines changed: 350 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,350 @@
1+
---
2+
title: Hypermode Graph MCP
3+
description:
4+
"Connect your Dgraph database to AI coding assistants using the Model Context
5+
Protocol for exploratory data analysis and intelligent graph operations"
6+
mode: "wide"
7+
sidebarTitle: "Graph MCP"
8+
"og:title": "Graph MCP"
9+
---
10+
11+
## Overview
12+
13+
The Hypermode Graph MCP (Model Context Protocol) server enables seamless
14+
integration between your graph database and AI coding assistants like Claude
15+
Desktop, Cursor, a nd other MCP-compatible tools. Two MCP endpoints are
16+
available with common tools for AI coding assistants: `mcp` – a server that
17+
provides tools and data from your Dgraph cluster · `mcp-ro` – a server that
18+
provides tools and data from your Dgraph cluster in read-only mode
19+
20+
<Note>
21+
**What is MCP?** The Model Context Protocol is an open standard that enables
22+
developers to build secure, two-way connections between their data sources and
23+
AI-powered tools. Think of it as a universal connector that allows AI
24+
assistants to understand and interact with your graph data.
25+
</Note>
26+
27+
## Why use Hypermode Graph MCP?
28+
29+
### For exploratory data analysis
30+
31+
- **Interactive graph exploration**: Query your knowledge graphs using natural
32+
language through your AI assistant
33+
- **Pattern Discovery**: Let AI help identify relationships, clusters, and
34+
anomalies in your graph data
35+
- **Dynamic schema understanding**: AI assistants can introspect your graph
36+
structure and suggest optimal queries
37+
- **Real-time insights**: Get immediate answers about your data without writing
38+
complex DQL queries
39+
40+
### For AI coding assistants
41+
42+
- **Context-aware development**: Your coding assistant understands your graph
43+
schema and data patterns
44+
- **Intelligent query generation**: AI can write and optimize DQL queries based
45+
on your specific use case
46+
- **Schema evolution support**: Get suggestions for schema changes and
47+
migrations
48+
- **Debugging assistance**: AI can help troubleshoot graph queries and
49+
performance issues
50+
51+
## Architecture
52+
53+
The Dgraph MCP server follows the standard MCP architecture:
54+
55+
```mermaid
56+
graph LR
57+
A[AI Assistant<br/>Claude/Cursor] --> B[MCP Client]
58+
B --> C[Dgraph MCP Server]
59+
C --> D[Hypermode Graph]
60+
C --> E[Local Dgraph Cluster]
61+
```
62+
63+
**Components:**
64+
65+
- **MCP Host**: Your AI-powered application (Claude Desktop, IDE with AI
66+
features)
67+
- **MCP Client**: Protocol client that maintains connection with the Dgraph
68+
server
69+
- **Dgraph MCP Server**: Exposes graph capabilities through standardized MCP
70+
interface
71+
- **Graph Database**: Your Dgraph cluster (local or on Hypermode)
72+
73+
## MCP server types
74+
75+
Dgraph provides two MCP server variants to match your security and usage
76+
requirements:
77+
78+
### Standard MCP server (`mcp`)
79+
80+
- **Full Access**: Read and write operations on your graph
81+
- **Use Cases**: Development environments, data modeling, schema evolution
82+
- **Capabilities**: Query execution, mutations, schema modifications, namespace
83+
management
84+
85+
### Read-only MCP server (`mcp-ro`)
86+
87+
- **Safe Exploration**: Query-only access to your graph data
88+
- **Use Cases**: Production analysis, reporting, exploratory data analysis
89+
- **Capabilities**: DQL queries, schema introspection, data visualization
90+
support
91+
92+
<Warning>
93+
Choose the read-only server (`mcp-ro`) for production environments or when
94+
working with sensitive data to prevent accidental modifications.
95+
</Warning>
96+
97+
## Setup guide
98+
99+
When using Hypermode Graphs, the MCP configuration is available on the graph
100+
details screen in the console:
101+
102+
<Steps>
103+
<Step title="Access Graph Console">
104+
Navigate to your Hypermode workspace and select your graph instance.
105+
</Step>
106+
107+
<Step title="Copy MCP Configuration">
108+
From the graph details screen, copy the provided MCP configuration.
109+
</Step>
110+
111+
<Step title="Configure Your AI Assistant">
112+
Add the configuration to your AI assistant's settings:
113+
114+
```json
115+
{
116+
"mcpServers": {
117+
"hypermode-graph": {
118+
"command": "npx",
119+
"args": [
120+
"mcp-remote",
121+
"https://<graph-workspace>.hypermode.host/mcp/sse",
122+
"--header",
123+
"Authorization: Bearer <bearer-token>"
124+
]
125+
}
126+
}
127+
}
128+
```
129+
130+
</Step>
131+
</Steps>
132+
133+
## Available Capabilities
134+
135+
```mermaid
136+
graph TD
137+
subgraph TopRow [" "]
138+
direction LR
139+
subgraph Client ["MCP Client"]
140+
C1["Invokes Tools"]
141+
C2["Queries for Resources"]
142+
C3["Interpolates Prompts"]
143+
end
144+
145+
subgraph Server ["MCP Server"]
146+
S1["Exposes Tools"]
147+
S2["Exposes Resources"]
148+
S3["Exposes Prompt Templates"]
149+
end
150+
151+
Client <--> Server
152+
end
153+
154+
subgraph BottomRow [" "]
155+
direction LR
156+
subgraph Tools ["Tools"]
157+
direction TB
158+
T1["Functions and tools that<br/>can be invoked by the<br/>client"]
159+
T2["Retrieve / search"]
160+
T3["Send a message"]
161+
T4["Update DB records"]
162+
end
163+
164+
subgraph Resources ["Resources"]
165+
direction TB
166+
R1["Read-only data or<br/>exposed by the server"]
167+
R2["Files"]
168+
R3["Database Records"]
169+
R4["API Responses"]
170+
end
171+
172+
subgraph Templates ["Prompt Templates"]
173+
direction TB
174+
P1["Pre-defined templates<br/>for AI interactions"]
175+
P2["Document Q&A"]
176+
P3["Transcript Summary"]
177+
P4["Output as JSON"]
178+
end
179+
end
180+
181+
TopRow --> BottomRow
182+
183+
classDef clientBox fill:#f9f9f9,stroke:#d4b896,stroke-width:2px,color:#000
184+
classDef serverBox fill:#e8dcc6,stroke:#d4b896,stroke-width:2px,color:#000
185+
classDef sectionHeader fill:#8b4513,color:#fff,font-weight:bold
186+
classDef description fill:#e8dcc6,stroke:#d4b896,stroke-width:1px,color:#000
187+
classDef example fill:#e8dcc6,stroke:#d4b896,stroke-width:1px,color:#000
188+
189+
class Client clientBox
190+
class Server serverBox
191+
class T1,R1,P1 description
192+
class T2,T3,T4,R2,R3,R4,P2,P3,P4 example
193+
```
194+
195+
### Tools
196+
197+
Interactive tools for graph operations:
198+
199+
The MCP servers provide the following tools:
200+
201+
#### `get_schema`
202+
203+
Retrieve the current schema of your graph database.
204+
205+
#### `run_query`
206+
207+
Run a DQL query on your graph database.
208+
209+
**Parameters:**
210+
211+
- `query` (string): DQL query to execute
212+
213+
#### `run_mutation`
214+
215+
Run a DQL mutation on your graph database.
216+
217+
**Parameters:**
218+
219+
- `mutation` (string): DQL mutation to execute
220+
221+
#### `alter_schema`
222+
223+
Modify the DQL schema of your graph database.
224+
225+
**Parameters:**
226+
227+
- `schema` (string): DQL schema to apply
228+
229+
#### `get_common_queries`
230+
231+
Provides reference queries to aide in query syntax.
232+
233+
### Resources
234+
235+
The MCP server exposes your graph data as accessible resources:
236+
237+
#### `dgraph://schema`
238+
239+
The current Dgraph DQL schema.
240+
241+
#### `dgraph://common_queries`
242+
243+
Pre-built query patterns for common operations.
244+
245+
### Prompts
246+
247+
Pre-configured prompt templates for common graph operations:
248+
249+
- **Data Exploration**: "Analyze the structure and relationships in this graph"
250+
- **Performance Optimization**: "Suggest improvements for this query
251+
performance"
252+
- **Schema Design**: "Help design an optimal schema for this data model"
253+
- **Migration Planning**: "Plan a migration strategy for this schema change"
254+
255+
#### `quickstart_prompt`
256+
257+
A quickstart prompt for getting started with graph MCP.
258+
259+
## Example workflows
260+
261+
### Exploratory data analysis
262+
263+
<iframe
264+
src="https://www.youtube.com/embed/cKJuzAs5Www"
265+
title="Hypermode Graph MCP With Claude Desktop"
266+
frameborder="0"
267+
allow="accelerometer; autoplay; clipboard-write;
268+
encrypted-media; gyroscope; picture-in-picture"
269+
allowfullscreen
270+
style={{ aspectRatio: "16 / 9", width: "100%" }}
271+
></iframe>
272+
273+
<Steps>
274+
<Step title="Understand your data">
275+
Ask your AI assistant: "What does my graph schema look like and what are the main entity types?"
276+
</Step>
277+
278+
<Step title="Discover Patterns">
279+
"Find all users who have more than 10 connections and show me their
280+
relationship patterns."
281+
</Step>
282+
283+
<Step title="Analyze Distributions">
284+
"What's the distribution of node types in my graph? Are there any outliers or
285+
interesting clusters?"
286+
</Step>
287+
288+
<Step title="Performance Insights">
289+
"Which queries are running slowly and how can I optimize them?"
290+
</Step>
291+
</Steps>
292+
293+
### AI-assisted development
294+
295+
<Steps>
296+
<Step title="Schema design">
297+
"I want to model a social network with users, posts, and interactions. What's the optimal schema design?"
298+
</Step>
299+
300+
<Step title="Query Optimization">
301+
"This query is slow: `[paste query]`. How can I improve its performance?"
302+
</Step>
303+
304+
<Step title="Index Recommendations">
305+
"Based on my query patterns, what indices should I add to improve
306+
performance?"
307+
</Step>
308+
309+
<Step title="Migration Planning">
310+
"I need to add a new 'tags' predicate to existing posts. What's the safest migration approach?"
311+
</Step>
312+
</Steps>
313+
314+
## Best practices
315+
316+
### Security
317+
318+
- **Use Read-Only Servers**: Default to `mcp-ro` for analysis and exploration
319+
- **Authentication**: Always use bearer tokens for Hypermode connections
320+
321+
### Development workflow
322+
323+
- **Start with Read-Only**: Begin exploration with `mcp-ro` to understand your
324+
data
325+
- **Iterative Schema Design**: Use AI assistance for gradual schema evolution
326+
- **Query Testing**: Validate AI-generated queries in development before
327+
production
328+
- **Documentation**: Keep schema documentation updated for better AI
329+
understanding
330+
331+
## Community and support
332+
333+
<CardGroup cols={3}>
334+
<Card title="Discord Community" icon="discord" href="https://discord.gg/hypermode">
335+
Join discussions about Graph MCP and get help from the community
336+
</Card>
337+
338+
<Card
339+
title="GitHub Issues"
340+
icon="github"
341+
href="https://github.com/hypermodeinc/dgraph/issues"
342+
>
343+
Report bugs and request features for the MCP server
344+
</Card>
345+
346+
<Card title="Documentation" icon="book" href="https://docs.hypermode.com/dgraph">
347+
Comprehensive Dgraph documentation and guides
348+
</Card>
349+
</CardGroup>
350+
```

0 commit comments

Comments
 (0)