DSL Starter Kit

Build parsers for your domain-specific languages without the pain.

---

What This Solves

You have a custom format you need to parse:

route GET /users/{id} -> UserController.show;

SELECT name FROM users WHERE age > 25

[database]
host = localhost
port = 5432

You don't want to write regex spaghetti. You don't want to spend weeks learning parser theory.

This kit lets you describe your language in plain English, and an AI generates a working parser.

---

How It Works

1. You describe your language to Claude (or Cursor)
2. The AI writes and validates the grammar
3. You test with sample inputs
4. The AI generates parser code in your language (Python, Java, TypeScript, etc.)

The magic: an MCP server gives the AI real parsing tools, so it validates instead of guessing.

---

Setup

1. Pull the Docker Image

docker pull sshailabh1/antlr4-mcp-server:latest

2. Configure MCP server in MCP Client(Claude Desktop/ChatGPT)

Claude Desktop

Add to your config file: claude_desktop_config.json

{
  "mcpServers": {
    "antlr4": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "sshailabh1/antlr4-mcp-server:latest"]
    }
  }
}

Start Claude Desktop after saving.

3. Verify Setup

Ask MCP CLient(AI assistant):

"List all available ANTLR4 tools"

You should see 9 tools listed.

---

Try It

Ask Claude:

"I need to parse config files like this:

[server]
port = 8080

Write a grammar, validate it, and parse my example."

Claude will:

• Write the grammar
• Validate it (catching real errors)
• Show you the parse tree
• Generate a parser in your language

---

Example Grammars Included

File	What It Parses
Hello.g4	Simple greetings (learning example)
Calculator.g4	Math expressions with precedence
RouteDsl.g4	HTTP route definitions
ConfigDsl.g4	INI-style config files
JsonSubset.g4	JSON documents
SqlSubset.g4	Basic SQL queries

Each has a matching sample file in samples/.

---

Generate a Parser

Once your grammar works:

"Generate a Python parser with visitor support."

Install the runtime and use it:

pip install antlr4-python3-runtime

from antlr4 import CommonTokenStream, InputStream
from RouteDslLexer import RouteDslLexer
from RouteDslParser import RouteDslParser

lexer = RouteDslLexer(InputStream(text))
parser = RouteDslParser(CommonTokenStream(lexer))
tree = parser.file_()

---

Supported Languages

Python, Java, JavaScript, TypeScript, Go, C#, C++, Swift, PHP, Dart

---

What the AI Can Do

Ask This	What Happens
"Validate this grammar"	Checks for syntax errors
"Parse this input"	Shows the parse tree
"Check for ambiguities"	Finds parsing conflicts
"Generate a Python parser"	Creates working code
"Show the rule dependencies"	Visualizes grammar structure

---

License

Apache 2.0