Live & Deployed on Hugging Face Spaces

Codebase Memory
MCP Server

Model Context Protocol server providing AI agents with deep knowledge of the Evolva codebase — 54 repositories, 7 architecture layers, full dependency graph.

54
Repos Indexed
6
MCP Tools
7
Arch Layers
17
Relationships
🚀 Open Space 📖 Quickstart 🔌 API Reference

6 Powerful Tools

Every tool is callable via the MCP protocol or directly via REST POST to /mcp/call.

🔍
search_repos

Search repositories by name, topic, language, or architecture layer with relevance scoring.

query* layer language limit
📦
get_repo_details

Get detailed information about a specific repository including all metadata and dependency relationships.

repo_name*
🏗️
get_architecture

Get the full 7-layer architecture overview with all repositories organized by layer.

layer_id
🔗
get_dependencies

Get the dependency graph for a repository — what it depends on (upstream) and what depends on it (downstream).

repo_name* direction
📊
get_stats

Get aggregate statistics: language distribution, layer sizes, star counts, and topic frequency analysis.

metric
🌐
find_related

Find repositories related to a given repo based on shared topics, language, layer proximity, or dependency graph.

repo_name* strategy limit

7-Layer Architecture

54 repositories organized into 7 logical layers from infrastructure to research.

L1

Foundation & Infrastructure

Core infrastructure, DevOps, CI/CD, containerization, and base platform services

7 repos
L2

Data & Storage Layer

Databases, vector stores, knowledge graphs, and data persistence systems

7 repos
L3

AI & ML Core

Machine learning models, training pipelines, inference engines, and AI primitives

7 repos
L4

Agent Framework

Agent orchestration, multi-agent systems, memory management, and tool use

7 repos
L5

MCP & Integration Layer

Model Context Protocol servers, API integrations, and connector ecosystem

9 repos
L6

Applications & Services

End-user applications, web services, dashboards, and product interfaces

7 repos
L7

Research & Experimental

Research projects, experimental features, prototypes, and innovation lab

8 repos

REST Endpoints

All endpoints are available at the Hugging Face Space URL.

MethodEndpointDescription
GET/healthHealth check — returns status and repos_indexed count
GET/mcp/toolsList all 6 MCP tool definitions
POST/mcp/callExecute an MCP tool call with JSON body
GET/v1/reposList all 54 repositories (supports ?layer, ?language, ?topic filters)
GET/v1/repos/{name}Get details for a specific repository
GET/v1/architectureGet the 7-layer architecture (supports ?layer_id filter)
GET/v1/statsGet aggregate codebase statistics
GET/v1/relationshipsGet all dependency relationships
GET/docsThis documentation site
GET/api/docsInteractive Swagger UI

Get Started in 60 Seconds

Connect to the MCP server or call the REST API directly.

MCP Configuration (Claude Desktop / Cursor)
{
  "mcpServers": {
    "evolva-codebase-memory": {
      "url": "https://pippinlitli-evolva-codebase-memory.hf.space",
      "transport": "http"
    }
  }
}
Python — Search Repositories
import requests

BASE = "https://pippinlitli-evolva-codebase-memory.hf.space"

# Search for agent-related repos
resp = requests.post(f"{BASE}/mcp/call", json={
    "name": "search_repos",
    "arguments": {"query": "agent", "layer": 4}
})
print(resp.json())

# Get architecture overview
arch = requests.get(f"{BASE}/v1/architecture").json()
print(f"Layers: {arch['total_layers']}, Repos: {arch['total_repos']}")
cURL — Health Check
curl https://pippinlitli-evolva-codebase-memory.hf.space/health
# {"status":"up","repos_indexed":54,"version":"1.0.0","layers":7}

curl https://pippinlitli-evolva-codebase-memory.hf.space/mcp/tools | jq '.count'
# 6