nexus-ai-fs 0.7.0

AI-Native Distributed Filesystem Architecture

Nexus

Documentation • Quickstart • PyPI • Examples

The AI-native filesystem for cognitive agents.

⚠️ Beta: Nexus is under active development. APIs may change.

🎯 What is Nexus?

Nexus is a virtual filesystem server for AI agents. It unifies files, databases, APIs, and SaaS tools into a single path-based API with built-in permissions, memory, semantic search, and skills lifecycle management.

Humans manage context. Agents operate within it.

The Nexus Control Panel lets humans curate the files, memories, permissions, and integrations that agents can access—providing oversight and control while agents focus on execution.

This repo contains the open-source server, SDK, CLI, and examples.

Highlights

Hero Features

• Universal Connectors — One API for local files, S3/GCS, Gmail, Google Drive, X/Twitter, and custom MCP servers
• Sandboxed Execution — Run code safely with Docker/E2B integration
• Grep Everything — nexus grep and semantic search across all backends with connector cache
• Skills System — Package agent capabilities with versioning, governance, and exporters (LangGraph, MCP, CLI)

Core Platform

• Control Panel — User management, permissions, integrations, audit & versioning
• Runtime Modules — Workspaces, Memory, Workflows, Files, Skills, Sandbox
• ReBAC Permissions — Zanzibar-style access control with multi-tenant isolation

Advanced

• Event-driven workflows, semantic search + LLM document reading, content deduplication + versioning, 14 MCP tools, plugin system, batch operations

Design Principles

• Everything is a File — Memories, CRM records, MCP servers, and documents all expose the same interface with paths, metadata, and permissions for unified search and composability.
• Just-in-Time Retrieval — Instead of pre-indexing everything, Nexus retrieves exactly what's needed when it's needed, adapting the retrieval strategy to each query.

🧑‍💻 Getting Started

Local Server

git clone https://github.com/nexi-lab/nexus.git
cd nexus
cp .env.example .env

# Set env
# Required: ANTHROPIC_API_KEY
# Optional: TAVILY_API_KEY, FIRECRAWL_API_KEY, NEXUS_OAUTH_GOOGLE_CLIENT_ID, NEXUS_OAUTH_GOOGLE_CLIENT_SECRET

./scripts/local-demo.sh --start

Docker Server

./scripts/docker-demo.sh --init
./scripts/docker-demo.sh --start

Python SDK

pip install nexus-ai-fs

import nexus

nx = nexus.connect(config={
    "url": "http://localhost:2026",
    "api_key": "nxk_..."
})

# File operations
nx.write("/workspace/hello.txt", b"Hello, Nexus!")
print(nx.read("/workspace/hello.txt").decode())

# Search across all backends
results = nx.grep("TODO", "/workspace")

# Agent memory
nx.memory.store("User prefers detailed explanations", memory_type="preference")
context = nx.memory.query("What does the user prefer?")

nx.close()

CLI

pip install nexus-ai-fs

export NEXUS_URL="http://localhost:2026"
export NEXUS_API_KEY="nxk_..."

nexus write /workspace/hello.txt "Hello, Nexus!"
nexus cat /workspace/hello.txt
nexus grep "TODO" /workspace
nexus memory store "Important fact" --type fact

💡 Why Nexus?

Problem	Nexus Solution
Fragmented APIs	One path-based API for all connectors
Permission chaos	Zanzibar-style ReBAC with audit trails
Ephemeral memory	Persistent memory with semantic retrieval
Agent silos	Shared workspaces and skills registry
No oversight	Control Panel for human-in-the-loop management
Ad-hoc tools	Skills lifecycle management with versioning, governance, and multi-format export

🔧 How Nexus Works

Access Layer: Agents connect via CLI, MCP Server, Python SDK, or REST API.

Application Layer:

• Control Panel — Humans manage users, permissions, integrations, and audit logs
• Runtime — Core operations (glob, grep, semantic search, file ops, run code) and modules (workspaces, memory, workflows, files, skills, sandbox)

Infrastructure Layer: OAuth, ReBAC permissions engine, FUSE mounting, RAG pipelines.

Cache Layer: L1 in-memory cache and L2 PostgreSQL cache for vector embeddings, ReBAC tuples, and metadata.

Connectors: X/Twitter, MCP servers, Gmail, Local FS, GCS, Google Drive, AWS S3, and custom connectors.

📚 Examples

Framework	Description	Location
CrewAI	Multi-agent collaboration	examples/crewai/
LangGraph	Permission-based workflows	examples/langgraph_integration/
Claude SDK	ReAct agent pattern	examples/claude_agent_sdk/
OpenAI Agents	Multi-tenant with memory	examples/openai_agents/
CLI	40+ shell demos	examples/cli/

📖 Documentation

• Installation
• Quick Start
• Control Panel Guide
• API Reference
• Skills System
• Multi-Tenant Architecture
• Permissions & ReBAC

🤝 Contributing

git clone https://github.com/nexi-lab/nexus.git
cd nexus
pip install -e ".[dev]"
pre-commit install
pytest tests/

See CONTRIBUTING.md for guidelines.

📝 License

© 2025 Nexi Labs, Inc. Licensed under Apache License 2.0 - See LICENSE for details.

---

If Nexus helps your project, please ⭐ the repo!