NeuralDrive

AI-native workspace with vector-indexed storage, context-aware document generation, and agentic workflows.

Features

Document Drive — Upload and manage documents (PDF, DOCX, CSV, XLSX, TXT) with automatic parsing, chunking, and vector indexing. Organize files in a hierarchical folder system.
Copilot — Conversational AI assistant with selectable document context, knowledge hub scoping, and MCP tool integration. Supports streaming responses with source citations.
Semantic Search — Search across all documents using natural language. Hybrid search combines vector similarity with keyword matching via pgvector.
Document Factory — AI-powered content generation with RAG context from your knowledge base. Includes templates, summarization, and Q&A.
Automations — Visual workflow builder with event-driven triggers (file upload, schedule, webhook, manual) and configurable action nodes. Supports conditional logic, loops, and parallel execution.
MCP Tools — Connect external tools and services via the Model Context Protocol. Register STDIO or HTTP servers, discover their capabilities, and use them in Copilot and workflows.
Knowledge Hubs — Curated document collections with dedicated chat interfaces. Configure RAG settings per hub with auto-inclusion rules.
AI Tables — Extract structured data from documents into configurable table schemas. Define columns with extraction prompts and export to CSV/JSON.

Quick Start

Using Docker (Recommended)

Clone the repository and copy the environment file:

cp .env.example .env

Add your API keys to .env:

OPENAI_API_KEY=your-key-here
# or
ANTHROPIC_API_KEY=your-key-here

Start all services:

docker-compose up --build

Access the application:

Frontend: http://localhost:3000
Backend API: http://localhost:8000
API Docs: http://localhost:8000/docs

Local Development

Backend

cd backend

# Create virtual environment
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows

# Install dependencies
pip install -r requirements.txt

# Start PostgreSQL and Redis (via Docker or locally)
docker-compose up postgres redis -d

# Run database migrations
alembic upgrade head

# Start the server
uvicorn app.main:app --reload

Frontend

cd frontend

# Install dependencies
npm install

# Start development server
npm run dev

Celery Workers

cd backend

# Worker (background tasks)
celery -A app.worker.celery_app worker --loglevel=info

# Beat (scheduled tasks)
celery -A app.worker.celery_app beat --loglevel=info

Architecture

NeuralDrive uses a hexagonal architecture (ports & adapters) on the backend and Next.js App Router on the frontend.

DMS/
├── backend/                      # FastAPI backend (Python 3.11)
│   ├── app/
│   │   ├── api/v1/              # Route handlers
│   │   │   ├── auth.py          # Authentication
│   │   │   ├── documents.py     # Document CRUD & upload
│   │   │   ├── folders.py       # Folder management
│   │   │   ├── search.py        # Semantic & hybrid search
│   │   │   ├── generation.py    # Content generation & RAG
│   │   │   ├── automations.py   # Workflow automation
│   │   │   ├── mcp.py           # MCP server management
│   │   │   ├── ai_tables.py     # AI table operations
│   │   │   ├── knowledge_hubs.py# Knowledge hub management
│   │   │   └── hub_chat.py      # Knowledge hub chat
│   │   │
│   │   ├── domain/              # Core business logic
│   │   │   ├── entities/        # Domain models (dataclasses)
│   │   │   ├── ports/           # Abstract interfaces (LLM, embedding, storage)
│   │   │   └── events/          # Domain events
│   │   │
│   │   ├── application/         # Use cases & services
│   │   │   └── services/        # Business logic orchestration
│   │   │       ├── agent_service.py       # Unified agent executor
│   │   │       ├── agentic_rag_service.py # Two-stage RAG retrieval
│   │   │       ├── ingestion_service.py   # Document processing
│   │   │       ├── search_service.py      # Search operations
│   │   │       ├── generation_service.py  # LLM generation
│   │   │       ├── automation_service.py  # Workflow management
│   │   │       ├── mcp_service.py         # MCP integration
│   │   │       ├── tool_registry.py       # Tool discovery & management
│   │   │       ├── ai_table_service.py    # AI table logic
│   │   │       ├── knowledge_hub_service.py
│   │   │       └── chat_service.py
│   │   │
│   │   ├── infrastructure/      # External implementations
│   │   │   ├── adapters/        # Port implementations
│   │   │   │   ├── llm/         # OpenAI, Anthropic, Ollama
│   │   │   │   ├── embedding/   # OpenAI, local sentence-transformers
│   │   │   │   ├── storage/     # Local filesystem, S3
│   │   │   │   └── web_search/  # Tavily
│   │   │   ├── database/        # SQLAlchemy models & repositories
│   │   │   └── parsers/         # PDF, DOCX, CSV, XLSX, TXT parsers
│   │   │
│   │   ├── worker/              # Celery background tasks
│   │   │   └── tasks/           # Ingestion, embedding, automation, AI table tasks
│   │   │
│   │   ├── config.py            # Pydantic settings
│   │   ├── dependencies.py      # FastAPI dependency injection
│   │   └── main.py              # App factory & lifespan
│   │
│   ├── alembic/                 # Database migrations
│   └── tests/                   # pytest test suite
│
├── frontend/                    # Next.js 14 frontend (TypeScript)
│   └── src/
│       ├── app/                 # App Router pages
│       │   ├── copilot/         # AI assistant
│       │   ├── drive/           # File manager
│       │   ├── editor/          # Document editor
│       │   ├── search/          # Search interface
│       │   ├── automations/     # Workflow management & builder
│       │   ├── mcp/             # MCP server configuration
│       │   ├── ai-tables/       # AI table interface
│       │   ├── knowledge-hubs/  # Knowledge hub management & chat
│       │   └── settings/        # Application settings
│       ├── components/          # React components
│       │   ├── ui/              # Base UI components
│       │   ├── copilot/         # Copilot interface
│       │   ├── workflow/        # Workflow builder (ReactFlow)
│       │   ├── ai-tables/       # AI table components
│       │   ├── knowledge-hubs/  # Knowledge hub components
│       │   ├── drive/           # File browser
│       │   ├── editor/          # Rich text editor (TipTap)
│       │   └── settings/        # Settings & MCP config
│       ├── lib/api/             # API client modules
│       ├── stores/              # Zustand state stores
│       ├── hooks/               # Custom React hooks
│       └── types/               # TypeScript type definitions
│
└── docker-compose.yml           # PostgreSQL, Redis, Backend, Celery, Frontend

API Reference

All endpoints are prefixed with /api/v1. Full interactive docs available at /docs when running.

Authentication (`/auth`)

Method	Endpoint	Description
POST	`/auth/register`	Register new user
POST	`/auth/login`	Login and get JWT tokens
POST	`/auth/refresh`	Refresh access token
GET	`/auth/me`	Get current user profile

Documents (`/documents`)

Method	Endpoint	Description
POST	`/documents/upload`	Upload a document file
POST	`/documents/create`	Create a text document
GET	`/documents/`	List documents
GET	`/documents/{id}`	Get document details
PATCH	`/documents/{id}/content`	Update document content
GET	`/documents/{id}/download`	Download document file
DELETE	`/documents/{id}`	Delete a document
POST	`/documents/{id}/reprocess`	Reprocess a document

Folders (`/folders`)

Method	Endpoint	Description
POST	`/folders/`	Create a folder
GET	`/folders/`	List folders at level
GET	`/folders/{id}`	Get folder with contents
GET	`/folders/tree/all`	Get full folder tree
GET	`/folders/{id}/breadcrumb`	Get breadcrumb path
PATCH	`/folders/{id}`	Update a folder
DELETE	`/folders/{id}`	Delete a folder
POST	`/folders/{id}/move`	Move a folder

Search (`/search`)

Method	Endpoint	Description
GET	`/search/?q=query`	Semantic search
GET	`/search/hybrid?q=query`	Hybrid search (vector + keyword)
GET	`/search/context?q=query`	Get RAG context

Generation (`/generation`)

Method	Endpoint	Description
POST	`/generation/generate`	Generate content with RAG
POST	`/generation/generate/agentic`	Agentic RAG with tool use
POST	`/generation/generate/stream`	Stream generated content
POST	`/generation/template`	Generate from template
POST	`/generation/summarize`	Summarize a document
POST	`/generation/ask`	Answer a question with RAG

Automations (`/automations`)

Method	Endpoint	Description
POST	`/automations/`	Create automation
GET	`/automations/`	List automations
GET	`/automations/{id}`	Get automation
PATCH	`/automations/{id}`	Update automation
DELETE	`/automations/{id}`	Delete automation
POST	`/automations/{id}/pause`	Pause automation
POST	`/automations/{id}/resume`	Resume automation
POST	`/automations/{id}/trigger`	Manually trigger automation
GET	`/automations/{id}/runs`	Get run history
GET	`/automations/meta/trigger-types`	List trigger types
GET	`/automations/meta/action-types`	List action types

MCP Servers (`/mcp`)

Method	Endpoint	Description
POST	`/mcp/servers`	Register MCP server
GET	`/mcp/servers`	List MCP servers
GET	`/mcp/servers/{id}`	Get MCP server
PATCH	`/mcp/servers/{id}`	Update MCP server
DELETE	`/mcp/servers/{id}`	Delete MCP server
POST	`/mcp/servers/{id}/connect`	Connect and discover tools
POST	`/mcp/servers/{id}/disconnect`	Disconnect from server
GET	`/mcp/servers/{id}/tools`	Get server tools
POST	`/mcp/servers/{id}/tools/{name}`	Execute a tool

AI Tables (`/ai-tables`)

Method	Endpoint	Description
POST	`/ai-tables/`	Create AI table
GET	`/ai-tables/`	List AI tables
GET	`/ai-tables/{id}`	Get AI table
PUT	`/ai-tables/{id}`	Update AI table
DELETE	`/ai-tables/{id}`	Delete AI table
POST	`/ai-tables/{id}/columns`	Add column
PUT	`/ai-tables/{id}/columns/{col_id}`	Update column
DELETE	`/ai-tables/{id}/columns/{col_id}`	Delete column
GET	`/ai-tables/{id}/rows`	List rows
POST	`/ai-tables/{id}/rows`	Add row manually
POST	`/ai-tables/{id}/rows/process`	Extract data from documents
PUT	`/ai-tables/{id}/rows/{row_id}`	Update row value
POST	`/ai-tables/{id}/rows/{row_id}/review`	Mark row reviewed
GET	`/ai-tables/{id}/export`	Export table (CSV/JSON)

Knowledge Hubs (`/knowledge-hubs`)

Method	Endpoint	Description
POST	`/knowledge-hubs/`	Create hub
GET	`/knowledge-hubs/`	List hubs
GET	`/knowledge-hubs/{id}`	Get hub
PUT	`/knowledge-hubs/{id}`	Update hub
DELETE	`/knowledge-hubs/{id}`	Delete hub
POST	`/knowledge-hubs/{id}/documents`	Add documents
DELETE	`/knowledge-hubs/{id}/documents`	Remove documents
POST	`/knowledge-hubs/{id}/folders`	Add folders
DELETE	`/knowledge-hubs/{id}/folders`	Remove folders
POST	`/knowledge-hubs/{id}/rules`	Add inclusion rule
DELETE	`/knowledge-hubs/{id}/rules/{idx}`	Remove rule
GET	`/knowledge-hubs/{id}/documents`	List hub documents
POST	`/knowledge-hubs/{id}/refresh`	Refresh hub index

Knowledge Hub Chat (`/knowledge-hubs/{hub_id}/chat`)

Method	Endpoint	Description
POST	`/chat/sessions`	Create chat session
GET	`/chat/sessions`	List sessions
GET	`/chat/sessions/{id}`	Get session with messages
PATCH	`/chat/sessions/{id}`	Update session
DELETE	`/chat/sessions/{id}`	Delete session
POST	`/chat/sessions/{id}/archive`	Archive session
POST	`/chat/sessions/{id}/messages`	Send message
POST	`/chat/sessions/{id}/messages/stream`	Stream message (SSE)
GET	`/chat/sessions/{id}/messages`	Get messages

Configuration

LLM Providers

NeuralDrive supports multiple LLM providers via a pluggable adapter pattern. Set DEFAULT_LLM_PROVIDER in .env:

Provider	Models	Config
OpenAI	GPT-4, GPT-4 Turbo, GPT-3.5 Turbo	`OPENAI_API_KEY`, `OPENAI_MODEL`
Anthropic	Claude 3 Opus, Sonnet, Haiku	`ANTHROPIC_API_KEY`, `ANTHROPIC_MODEL`
Ollama	Llama 2, Mistral, and other local models	`OLLAMA_BASE_URL`, `OLLAMA_MODEL`

Embedding Providers

Set DEFAULT_EMBEDDING_PROVIDER in .env:

Provider	Models	Dimensions
OpenAI	text-embedding-3-small, text-embedding-3-large	1536
Local	all-MiniLM-L6-v2 (sentence-transformers)	384

Storage Backends

Set STORAGE_BACKEND in .env:

Backend	Config
Local	`STORAGE_PATH` (default: `./storage`)
S3	`S3_BUCKET`, `S3_REGION`, `S3_ACCESS_KEY`, `S3_SECRET_KEY`

Web Search (Optional)

Set ENABLE_WEB_SEARCH=true and configure TAVILY_API_KEY to enable web search in Copilot and agent workflows.

Authentication

JWT-based authentication. Configure SECRET_KEY, ALGORITHM (default: HS256), and ACCESS_TOKEN_EXPIRE_MINUTES (default: 30).

Tech Stack

Backend

FastAPI — async web framework
SQLAlchemy 2.0 — async ORM
PostgreSQL + pgvector — database with vector similarity search
Celery + Redis — background task processing and scheduling
Alembic — database migrations
Pydantic v2 — settings and request/response validation
passlib + python-jose — authentication (JWT + bcrypt)

Frontend

Next.js 14 — React framework (App Router)
TypeScript — type safety
TanStack Query — data fetching and caching
Zustand — state management (with Immer)
TipTap — rich text editor
ReactFlow — workflow visualization
Tailwind CSS — styling
React Dropzone — file uploads

Infrastructure

Docker Compose — orchestration (PostgreSQL, Redis, Backend, Celery Worker, Celery Beat, Frontend)

License

MIT

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
backend		backend
docs/screenshots		docs/screenshots
frontend		frontend
.env.example		.env.example
.gitignore		.gitignore
AGENTS.md		AGENTS.md
CLAUDE.md		CLAUDE.md
README.md		README.md
brief.md		brief.md
docker-compose.yml		docker-compose.yml
sample-tech.md		sample-tech.md

Folders and files

Latest commit

History

Repository files navigation

NeuralDrive

Features

Quick Start

Using Docker (Recommended)

Local Development

Backend

Frontend

Celery Workers

Architecture

API Reference

Authentication (/auth)

Documents (/documents)

Folders (/folders)

Search (/search)

Generation (/generation)

Automations (/automations)

MCP Servers (/mcp)

AI Tables (/ai-tables)

Knowledge Hubs (/knowledge-hubs)

Knowledge Hub Chat (/knowledge-hubs/{hub_id}/chat)

Configuration

LLM Providers

Embedding Providers

Storage Backends

Web Search (Optional)

Authentication

Tech Stack

Backend

Frontend

Infrastructure

License

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Authentication (`/auth`)

Documents (`/documents`)

Folders (`/folders`)

Search (`/search`)

Generation (`/generation`)

Automations (`/automations`)

MCP Servers (`/mcp`)

AI Tables (`/ai-tables`)

Knowledge Hubs (`/knowledge-hubs`)

Knowledge Hub Chat (`/knowledge-hubs/{hub_id}/chat`)

Packages