django-apcore

Django integration for the apcore protocol — scan your existing Django REST APIs and serve them as MCP tools for AI agents.

Overview

django-apcore bridges your existing Django REST Framework and django-ninja endpoints to the apcore protocol, enabling them to be served as MCP (Model Context Protocol) tools that AI agents can discover and invoke.

The core philosophy is scan, don't rewrite: instead of manually defining MCP tools alongside your API endpoints, django-apcore auto-scans your existing OpenAPI schemas (via drf-spectacular or django-ninja) and generates apcore module definitions. These modules are then served to AI agents through apcore-mcp.

Key Features

• DjangoApcore unified entry point — single class for all django-apcore functionality
• Auto-scan DRF endpoints via drf-spectacular OpenAPI generation
• Auto-scan django-ninja endpoints via built-in OpenAPI schema extraction
• Auto-resolve $ref — Pydantic model schemas resolved from OpenAPI $ref references
• Annotation inference — GET→readonly/cacheable, DELETE→destructive, PUT→idempotent
• Semantic module IDs — action verbs from function names (list, get, create) instead of raw HTTP methods
• Three output formats — YAML bindings, Python @module wrappers, or direct registry registration
• Output verification — validate generated files for syntax and structure
• AI enhancement — auto-enhance module metadata via local SLMs (Ollama/vLLM)
• Serve as MCP tools via apcore-mcp (stdio / streamable-http / SSE transports)
• Output formatting — Markdown or custom formatters for LLM-friendly results
• Pluggable middleware pipeline — logging, tracing, metrics, and custom middleware
• YAML-based access control (ACL) for fine-grained module permissions
• Django context factory — maps request.user to apcore Identity automatically
• Embedded MCP server mode — start MCP server alongside Django on startup
• Include/exclude endpoint filtering with regex patterns
• Export to OpenAI tool format for non-MCP integrations
• Tool Explorer — browser-based UI for browsing schemas and testing tools interactively
• JWT authentication — Bearer token auth on the MCP server

How It Works

Django Endpoints (DRF / django-ninja)
        │
        ▼
   ┌─────────┐     ┌───────────────┐
   │ Scanner  │────▶│ ScannedModule │  ← annotations auto-inferred
   └─────────┘     └───────┬───────┘
                           │
              ┌────────────┼────────────┐
              ▼            ▼            ▼
     YAML bindings   Python @module   Registry
              │            │            │  (direct)
              └────────────┴────────────┘
                           ▼
                    ┌──────────┐
                    │ Registry │
                    └────┬─────┘
                         ▼
                    ┌──────────┐    ┌─────────────┐
                    │ Executor │───▶│ Middlewares  │
                    └────┬─────┘    │ ACL / Tracing│
                         │         └─────────────┘
                         ▼
                    ┌───────────┐
                    │ apcore-mcp│
                    └────┬──────┘
                         ▼
                   MCP Transport
              (stdio / HTTP / SSE)
                         │
                         ▼
                    AI Agent

Quick Start

1. Install

pip install django-apcore[all]

2. Add to Django settings

INSTALLED_APPS = [
    # ...
    "django_apcore",
]

APCORE_MODULE_DIR = "apcore_modules/"

3. Scan your endpoints

# Generate YAML bindings from django-ninja routes
python manage.py apcore_scan --source ninja --output yaml

# Or register directly into the registry (no files)
python manage.py apcore_scan --source ninja --output registry

# Scan DRF endpoints
python manage.py apcore_scan --source drf --output yaml

# Preview without writing
python manage.py apcore_scan --source ninja --dry-run

# Filter with regex
python manage.py apcore_scan --source drf --include "users.*" --exclude "admin.*"

4. Serve as MCP tools

# Start MCP server (stdio, default)
python manage.py apcore_serve

# HTTP transport with explorer UI
python manage.py apcore_serve --transport streamable-http --port 9090 --explorer

DjangoApcore — Unified Entry Point

DjangoApcore is the recommended way to use django-apcore. It provides a single class for all functionality, following the same pattern as apcore.APCore and apcore_mcp.APCoreMCP.

from django_apcore import DjangoApcore

app = DjangoApcore()

Register modules

from apcore import ModuleAnnotations

@app.module(
    id="analytics.task_stats",
    tags=["analytics"],
    annotations=ModuleAnnotations(readonly=True, cacheable=True),
)
def task_stats() -> dict:
    """Return summary statistics about all tasks."""
    return {"total": 42, "done": 10, "pending": 32}

Call modules from Django views

from django.http import JsonResponse
from django_apcore import DjangoApcore

app = DjangoApcore()

# Sync view — request auto-maps to apcore Identity
def my_view(request):
    result = app.call("analytics.task_stats", request=request)
    return JsonResponse(result)

# Async view
async def my_async_view(request):
    result = await app.call_async("analytics.task_stats", request=request)
    return JsonResponse(result)

# Streaming
async def my_stream_view(request):
    async for chunk in app.stream("ai.chat", {"prompt": "hello"}, request=request):
        yield chunk

# With cancellation timeout
result = app.cancellable_call("slow.module", timeout=30, request=request)

Scan, list, and describe

# Scan endpoints programmatically
modules = app.scan(source="ninja", include="users")

# List registered modules
ids = app.list_modules(tags=["analytics"])

# Get module description (for LLM use)
desc = app.describe("analytics.task_stats")

Task management

# Submit async task
task_id = await app.submit_task("heavy.report", {"year": 2026})

# Check status
status = app.get_task_status(task_id)

# Cancel
await app.cancel_task(task_id)

MCP serving and export

# Start MCP server (blocking)
app.serve(transport="streamable-http", port=9090, explorer=True)

# Export as OpenAI tools
tools = app.to_openai_tools(tags=["public"], strict=True)

Singleton access

# Process-wide singleton (for use across modules)
app = DjangoApcore.get_instance()

Properties

Property	Returns	Description
app.registry	Registry	apcore Registry singleton
app.executor	Executor	apcore Executor with extensions applied
app.extension_manager	ExtensionManager	Extension point manager
app.context_factory	ContextFactory	Django request → apcore Context
app.metrics_collector	MetricsCollector | None	Metrics (if enabled)
app.task_manager	AsyncTaskManager	Async task manager
app.settings	ApcoreSettings	Validated Django settings

Installation

Install with pip, choosing the extras you need:

# All optional dependencies
pip install django-apcore[all]

# Only django-ninja scanner
pip install django-apcore[ninja]

# Only DRF scanner (via drf-spectacular)
pip install django-apcore[drf]

# Only MCP server support
pip install django-apcore[mcp]

Dependencies

Package	Version	Purpose
django	>= 4.2	Core framework
apcore	>= 0.13.0	Protocol SDK
apcore-toolkit	>= 0.2.0	Scanner, writer, and formatting utilities
pyyaml	>= 6.0	YAML binding files

Optional Dependencies

Extra	Package	Version	Purpose
ninja	django-ninja	>= 1.0	django-ninja endpoint scanning
drf	drf-spectacular	>= 0.27	DRF endpoint scanning via OpenAPI
mcp	apcore-mcp	>= 0.10.0	MCP server, transport layer, JWT auth, Tool Explorer

Configuration

All settings are prefixed with APCORE_ and read from Django's settings.py.

Core Settings

Setting	Type	Default	Description
APCORE_MODULE_DIR	str	"apcore_modules/"	Directory for YAML binding files and @module Python files
APCORE_AUTO_DISCOVER	bool	True	Auto-discover bindings and @module functions on Django startup
APCORE_BINDING_PATTERN	str	"*.binding.yaml"	Glob pattern for discovering YAML binding files
APCORE_CONTEXT_FACTORY	str | None	None	Dotted path to custom ContextFactory class
APCORE_EXECUTOR_CONFIG	dict | None	None	Additional executor configuration dict
APCORE_VALIDATE_INPUTS	bool	False	Enable input validation at the executor layer

MCP Server Settings

Setting	Type	Default	Description
APCORE_SERVE_TRANSPORT	str	"stdio"	MCP transport: "stdio", "streamable-http", or "sse"
APCORE_SERVE_HOST	str	"127.0.0.1"	Host for HTTP-based transports
APCORE_SERVE_PORT	int	9090	Port for HTTP-based transports (1–65535)
APCORE_SERVER_NAME	str	"apcore-mcp"	MCP server name (1–100 characters)
APCORE_SERVER_VERSION	str | None	None	Server version string
APCORE_OUTPUT_FORMATTER	str | None	None	Dotted path to output formatter (e.g., "apcore_toolkit.to_markdown")

Extension Settings

Setting	Type	Default	Description
APCORE_MIDDLEWARES	list[str]	[]	Dotted paths to middleware classes
APCORE_ACL_PATH	str | None	None	Path to YAML ACL file for access control
APCORE_MODULE_VALIDATORS	list[str]	[]	Dotted paths to extra module validator classes

Explorer Settings

Setting	Type	Default	Description
APCORE_EXPLORER_ENABLED	bool	False	Enable the browser-based Tool Explorer UI
APCORE_EXPLORER_PREFIX	str	"/explorer"	URL prefix for the explorer UI
APCORE_EXPLORER_ALLOW_EXECUTE	bool	False	Allow tool execution from the explorer UI

JWT Authentication Settings

Setting	Type	Default	Description
APCORE_JWT_SECRET	str | None	None	JWT secret/key. Enables JWT auth when set
APCORE_JWT_ALGORITHM	str	"HS256"	JWT algorithm (e.g., HS256, RS256)
APCORE_JWT_AUDIENCE	str | None	None	Expected JWT aud claim
APCORE_JWT_ISSUER	str | None	None	Expected JWT iss claim

AI Enhancement Settings

Setting	Type	Default	Description
APCORE_AI_ENHANCE	bool	False	Enable AI enhancement in apcore_scan

AI enhancement also requires the APCORE_AI_ENABLED=true environment variable and a running SLM (Ollama/vLLM). See apcore-toolkit for configuration.

Observability Settings

Setting	Type	Default	Description
APCORE_OBSERVABILITY_LOGGING	bool | dict | None	None	Enable observability logging middleware
APCORE_TRACING	bool | dict | None	None	Enable tracing (stdout, in_memory, otlp)
APCORE_METRICS	bool | dict | None	None	Enable metrics collection

Observability dict options

# Logging
APCORE_OBSERVABILITY_LOGGING = {
    "log_inputs": True,
    "log_outputs": True,
    "level": "info",          # trace, debug, info, warn, error, fatal
    "format": "json",
    "redact_sensitive": True,
}

# Tracing
APCORE_TRACING = {
    "exporter": "otlp",              # stdout, in_memory, otlp
    "sampling_rate": 0.1,            # 0.0 to 1.0
    "sampling_strategy": "full",     # full, proportional, error_first, off
    "otlp_endpoint": "http://localhost:4318",
    "otlp_service_name": "my-django-app",
}

# Metrics
APCORE_METRICS = {
    "buckets": [0.01, 0.05, 0.1, 0.25, 0.5, 1.0, 2.5, 5.0, 10.0],
}

Embedded Server Settings

Setting	Type	Default	Description
APCORE_EMBEDDED_SERVER	bool | dict | None	None	Start an embedded MCP server on Django startup

APCORE_EMBEDDED_SERVER = {
    "transport": "streamable-http",
    "host": "127.0.0.1",
    "port": 9090,
    "name": "embedded-mcp",
}

Task & Cancellation Settings

Setting	Type	Default	Description
APCORE_TASK_MAX_CONCURRENT	int	10	Max concurrent async tasks
APCORE_TASK_MAX_TASKS	int	1000	Max total tasks in queue
APCORE_TASK_CLEANUP_AGE	int	3600	Seconds before completed tasks are cleaned up
APCORE_CANCEL_DEFAULT_TIMEOUT	int | None	None	Default cancellation timeout (seconds)

Management Commands

apcore_scan

Scan Django API endpoints and generate apcore module definitions.

python manage.py apcore_scan --source <ninja|drf> [options]

Option	Short	Description
--source	-s	Required. Scanner source: ninja or drf
--output	-o	Output format: yaml (default), python, or registry
--dir	-d	Output directory (default: APCORE_MODULE_DIR)
--dry-run		Preview output without writing files
--verify		Verify written output files (YAML validity, Python syntax)
--include		Regex pattern to include endpoints
--exclude		Regex pattern to exclude endpoints
--ai-enhance		Enhance module metadata via local SLM

apcore_serve

Start an MCP server with registered apcore modules.

python manage.py apcore_serve [options]

Option	Short	Description
--transport	-t	stdio, streamable-http, or sse
--host		Host for HTTP transports
--port	-p	Port for HTTP transports
--name		MCP server name
--server-version		Server version string
--explorer		Enable the browser-based Tool Explorer UI (HTTP only)
--explorer-prefix		URL prefix for the explorer UI (default: /explorer)
--allow-execute		Allow tool execution from the explorer UI
--validate-inputs		Enable input validation
--metrics		Enable Prometheus /metrics endpoint
--log-level		Log level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
--tags		Filter modules by tags (comma-separated)
--prefix		Filter modules by ID prefix
--output-formatter		Dotted path to output formatter (e.g., apcore_toolkit.to_markdown)
--jwt-secret		JWT secret/key for authentication
--jwt-algorithm		JWT algorithm (default: HS256)
--jwt-audience		Expected JWT aud claim
--jwt-issuer		Expected JWT iss claim

apcore_export

Export registered modules to OpenAI tool format.

python manage.py apcore_export [options]

Option	Description
--format	Export format: openai-tools (default)
--strict	Enable strict mode for OpenAI Structured Outputs
--embed-annotations	Embed module annotations in output
--tags	Filter by tags (space-separated)
--prefix	Prefix for tool names

apcore_tasks

Manage async tasks.

python manage.py apcore_tasks <list|cancel|cleanup> [options]

Legacy API (Shortcuts)

For projects not using DjangoApcore, the function-based shortcuts are still available:

from django_apcore.shortcuts import (
    executor_call,          # sync call
    executor_call_async,    # async call
    executor_stream,        # streaming
    cancellable_call,       # call with CancelToken
    cancellable_call_async, # async cancellable
    submit_task,            # submit async task
    get_task_status,        # query task status
    cancel_task,            # cancel task
    report_progress,        # MCP progress
    elicit,                 # MCP user input
)

# Example
def my_view(request):
    result = executor_call("users.list", {"page": 1}, request=request)
    return JsonResponse(result)

Comparison with Other MCP Solutions

Feature	django-apcore	django-mcp-server	django-mcp	django-ninja-mcp	FastMCP	fastapi-mcp
Approach	Scan existing endpoints	Define new tools	Define new tools	Scan ninja endpoints	Define new tools	Scan FastAPI endpoints
DRF support	Yes (via drf-spectacular)	Yes (opt-in decorators)	No	No	No (not Django)	No (FastAPI only)
django-ninja support	Yes	No	No	Yes	No (not Django)	No (FastAPI only)
Unified client class	DjangoApcore	No	No	No	No	No
Schema source	Auto from OpenAPI	Model/serializer introspection	Manual	Auto from OpenAPI	Manual	Auto from OpenAPI
Transport: stdio	Yes	Yes	No	No	Yes	No
Transport: streamable-http	Yes	Yes	No	No	Yes	Yes
Transport: SSE	Yes	No	Yes	Yes	Yes	Yes
Annotation inference	Yes	No	No	No	No	No
Output verification	Yes	No	No	No	No	No
AI enhancement	Yes	No	No	No	No	No
Tracing (OpenTelemetry)	Yes	No	No	No	Yes	No
Metrics collection	Yes	No	No	No	No	No
Middleware pipeline	Yes	No	No	No	Yes	No
YAML-based ACL	Yes	No	No	No	No	No
Identity mapping	Yes (Django User)	DRF auth classes	No	No	Per-component auth	FastAPI Depends()
Export to OpenAI tools	Yes	No	No	No	No	No
JWT authentication	Yes	No	No	No	No	No

Demo Project

The example/ directory contains a Task Manager API demo showcasing all major features.

cd example
docker compose up --build

Service	URL	Description
web	http://localhost:8000	Django API server (django-ninja)
mcp	http://localhost:9090	MCP server with Tool Explorer

See example/README.md for full setup instructions and DjangoApcore usage examples.

Requirements

• Python 3.11+
• Django 4.2+ (including 5.0, 5.1, 6.0)

License

MIT