import { motion } from 'framer-motion'; import { Database, Upload, MessageSquare, Settings, Zap, Code, FileText, BookOpen } from 'lucide-react'; export default function HelpDocs() { return (
Everything you need to know about using Code-Forge
Click the + button in the sidebar to create a new KB. Choose the appropriate type for your content.
Go to KB Management tab and drag & drop files or click to browse. Supports text, markdown, PDF, code files, and more.
Switch to the Chat tab and ask questions about your documents. The AI will retrieve relevant context and provide answers.
General-purpose knowledge base for mixed content types.
Use for: Random notes, mixed documents, general reference material
Optimized for source code with better chunking and structure preservation.
Use for: Codebases, libraries, API implementations, code examples
Optimized for technical documentation with enhanced markdown processing.
Use for: MCP docs, API docs, tutorials, guides, technical specs
Special type for Agent OS YAML profiles with structured metadata extraction.
Use for: Agent OS configuration files only
Combines semantic vector search with keyword-based BM25 search for better retrieval.
When to use: When you need both semantic understanding AND exact keyword matching. Great for technical queries.
Re-scores retrieved documents using a cross-encoder model for higher accuracy.
When to use: When you need the most relevant results and can afford slightly slower queries. Improves precision.
Uses multi-step reasoning with an LLM agent to iteratively refine queries and answers.
When to use: For complex questions requiring multiple retrieval steps or reasoning. Slower but more intelligent.
💡 Pro Tip: Start with basic vector search, then enable Hybrid + Rerank for better results. Use Agentic RAG for complex research questions.
And many more! Code-Forge automatically extracts text from PDFs and processes markdown with enhanced structure preservation.
Code-Forge exposes 12 MCP tools for AI agents to interact with your knowledge bases:
list_knowledge_bases
create_knowledge_base
delete_knowledge_base
ingest_document
ingest_directory
query_knowledge_base
get_kb_stats
get_standards
MCP Endpoint: http://localhost:8051/mcp
Option 1: Using CLI (Recommended)
Run this command in your terminal:
npx @modelcontextprotocol/inspector http://localhost:8051/mcp
Or add directly to Claude Desktop:
claude mcp add code-forge http://localhost:8051/mcp
Option 2: Manual Configuration
Edit your Claude Desktop config file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{`{
"mcpServers": {
"code-forge": {
"url": "http://localhost:8051/mcp"
}
}
}`}
Note: Restart Claude Desktop after adding the configuration. The MCP server must be running (docker-compose up).