D2 Overview
D2 is a production-ready Python SDK that makes securing AI agents and LLM-powered tools straightforward and efficient. It replaces hundreds of lines of complex authorization logic with a single, powerful decorator. This allows developers to implement Role-Based Access Control (RBAC) at the code level, ensuring that functions are denied by default until explicitly authorized. D2 is designed for high-performance systems, offering both local file-based policies for development and a scalable cloud solution for production environments with minimal overhead.
How to use D2
Getting started with D2 is designed to be quick and simple, taking just a few minutes. First, install the SDK using pip: pip install d2-sdk. Next, protect your sensitive functions by adding the @d2.d2_guard("your_tool_name") decorator above them. After decorating your functions, run python -m d2 init in your terminal to automatically scan your code and generate a policy template. Finally, initialize D2 in your application's startup script and set the user context for each request, preferably using the with d2.set_user_context() manager to ensure proper handling.
Core Features of D2
- One Decorator Security: Secure any function by simply adding the
@d2_guarddecorator, instantly replacing complex auth logic. - Code-Level Security: Identifies and enforces authorization at the function level, not just API endpoints, with a deny-by-default security posture.
- Fast Policy Updates: Policy changes propagate across all environments in under a minute (as fast as 30 seconds on the Pro plan).
- Local & Cloud Modes: Start with simple local file-based policies for development and seamlessly scale to the cloud for production without code changes.
- Rich CLI Tools: A comprehensive command-line interface for policy management, inspection, and diagnostics to enhance the developer experience.
- Runtime Optimized: Built for high-performance production systems with intelligent caching and minimal overhead.
Use Cases for D2
D2 is ideal for developers building applications with AI agents or Large Language Models (LLMs) that interact with various tools and data sources. It can be used to secure access to internal APIs (e.g., a weather API), control actions that have external effects (e.g., sending emails or notifications), protect sensitive data operations (e.g., querying a user database), and manage permissions for file operations or interactions with third-party services.
Advantages of D2
The primary advantage of D2 is its simplicity and power. It dramatically reduces the amount of boilerplate code required for authorization, improving developer productivity and reducing the surface area for security bugs. Its runtime-optimized design ensures high performance, while the dual local/cloud modes provide a flexible workflow from development to production. The fast policy propagation allows for near-real-time permission changes, enhancing security and operational agility.
Pricing and Plans
D2 offers a freemium model with plans to suit different scales:
- Free Plan ($0): Perfect for development, includes 3 tools per app, 1 published app, 15-minute policy refresh, and 1 team member.
- Essentials Plan ($99/month): Designed for small teams, offers 10 tools per app, 2 published apps, 5-minute policy refresh, and 3 team members.
- Pro Plan ($249/month): The most popular option for growing applications, with 100 tools per app, 10 published apps, 30-second policy refresh, and 10 team members.
- Enterprise Plan (Custom): For large organizations requiring custom limits, features, and support.
All plans include file-based policies, CLI tools, community support, and a local development mode.
D2 FAQ
Traffic
Latest traffic
Status
Monthly traffic trend
- 2025-9: 461
- 2026-1: 0
- 2026-2: 0
- 2026-3: 177
- 2026-4: 0
- 2026-5: 296
Geography
Top 5 countries / regions
- 🇺🇸United States89.8%
- 🇮🇱Israel10.2%
Top keywords
| Keyword | Cost per click |
|---|---|
| artoo | $0.28 |
| rocoto | $0.80 |
| rocoto israel | $0.00 |
D2 Categories
D2 Jobs
D2 Embed Widget
Copy this embed code to place the badge on your blog, article, or product site and send readers directly to this ToolMage detail page.
Related AI tools
Oso
Oso is an Authorization as a Service platform for developers. It simplifies the implementation of complex access control logic like RBAC, ReBAC, and ABAC. Using its declarative policy language, Polar, engineering teams can quickly build and enforce fine-grained permissions for any application, including modern AI-native apps with agentic workflows and RAG systems, accelerating development and enhancing security.
Security
Morph
Morph is an API-first platform for developers, providing the fastest and most accurate way to apply AI-generated edits to code and files. With speeds of 4,500+ tokens/sec and 98% accuracy, it empowers AI agents to perform reliable, semantic modifications, bridging the gap between AI suggestions and production-ready implementation.
Infrastructure
Hoop.dev
Hoop.dev is an AI-powered access gateway providing developers with invisible security and admins with command-line control. It offers secure, auditable access to databases and servers, featuring real-time AI data masking, session recording, and streamlined approval workflows to enhance security without disrupting productivity.
Security
LangSearch
LangSearch provides free Web Search and Semantic Rerank APIs designed to connect LLM applications with clean, accurate, real-world context. It supports natural language queries, hybrid search, and offers a highly efficient reranker to improve result accuracy for AI agents, chatbots, and RAG systems.
Llm
Peargent
Peargent is a modern, powerful Python framework designed for building intelligent, production-grade AI agents. It offers an intuitive API, flexible LLM support, multi-agent orchestration, and persistent memory, enabling developers to create scalable and robust AI solutions for real-world use.
Agent Development
Golf
Golf is an enterprise-grade, protocol-aware firewall designed for the Model Context Protocol (MCP). It provides a centralized security layer to protect MCP servers from specific threats like prompt injection and token hijacking, enabling businesses to securely deploy AI agent infrastructure into production.
Agent Security
Cirtus AI
Cirtus AI is the trust layer for autonomous AI agents, providing decentralized identity, secure isolated wallets, and verifiable on-chain reputation. It empowers AI agents to act autonomously and securely, mitigating risks associated with direct access to sensitive assets and data, fostering a new era of trusted AI interactions.
Agent Frameworks
Thesys
Thesys provides C1, a Generative UI API that transforms LLM responses into dynamic, interactive user interfaces in real-time. Designed for developers building AI agents and applications, it automates frontend development by converting text outputs into live UI components like forms, charts, and tables. This allows for the creation of adaptive, engaging, and intelligent interfaces without the limitations of hardcoded UI, accelerating AI product development.
Api
Plandex
Plandex is an open-source, terminal-based AI coding agent designed for complex, large-scale software development tasks. It excels at handling large projects and files with its 2M token context window. Featuring a diff review sandbox, configurable autonomy, and multi-model support, Plandex empowers developers to build, debug, and refactor entire applications with AI assistance, ensuring control and productivity in a dev-friendly workflow.
Code Generation
Toolhouse
Toolhouse is a no-code/low-code platform that democratizes AI agent creation. It allows developers to build, deploy, and manage production-grade AI agents from simple prompts. With pre-built integrations for web scraping, search, RAG, and databases, Toolhouse abstracts away the complexity of infrastructure, enabling one-click deployment of agents as scalable APIs. It's designed to empower developers of all skill levels to ship AI-powered applications quickly.
Agent Development
InstaVM
InstaVM is a production-grade sandbox built for AI agents, offering hardware-isolated virtual machines with persistent state, secure networking, and secret management. It provides a complete Linux environment for safely executing untrusted code from agents, with sub-200ms cold starts and seamless deployment.
Virtual Machines
Syncline
Syncline is an AI-first scheduling infrastructure designed for developers and AI agents. It automates meeting scheduling by intelligently analyzing preferences and context, integrating seamlessly with Google Calendar via a powerful API and Model Context Protocol. It eliminates manual coordination, allowing AI agents to book optimal meeting times autonomously.
Api
Asimov
Asimov provides a foundational AI search API for developers to build intelligent agents and applications. It features built-in semantic search and re-ranking for high accuracy, simple content ingestion, and robust source management. The platform is designed with enterprise-grade security and offers detailed usage tracking, making it a comprehensive solution for creating custom search experiences.
Data Management
Permit.io
Permit.io is a full-stack authorization platform designed for the AI era. It simplifies the implementation of complex access controls like RBAC, ABAC, and ReBAC for developers. With a no-code policy editor, GitOps integration, and embeddable UI components, it allows entire teams to manage permissions securely and efficiently. The platform ensures low-latency decisions by running in a hybrid model, keeping sensitive data within your network while offering robust compliance and scalability for modern applications, including those powered by AI agents.
Security
oso.ai
oso.ai is an AI-powered authorization platform that helps developers build, manage, and enforce fine-grained access control. It simplifies complex security policies using natural language, intelligent automation, and a flexible policy engine.
Compliance
D2 Comments (0)
Sign in to comment.
Sign inNo comments yet.