Graph-Driven AI for All: Neo4j Aura Agent Enters General Availability

Our pioneering agent-creation platform, Neo4j Aura Agent, is now generally available—and free for the entire month of February 2026. 

AuraDB customers can now build & deploy knowledge graph grounded agents in minutes with additional powerful features – including automated, ontology driven agent construction and single click deployment to a hosted, secure MCP Server! 

Agentic AI isn’t just generating extraordinary buzz—it’s reshaping investment strategies worldwide. In the first half of 2025 alone, an estimated $2.8 billion flowed into the agentic space.

Data integration and “AI-ready” gaps are among the greatest challenges for enterprise agents. It isn’t just the cliché “fragmented data problem” – the real issue is lack of knowledgeable data integration at both agent runtime and construction.

At runtime, agents struggle to understand data models and execute the right tool calls. Teams also face challenges with secure deployment and connecting to the right data stores for retrieval. The second, often overlooked problem, is the agent construction and testing bottleneck. Building effective agents has two phases: creating an initial draft with domain-specific tools and prompts, then refining through repeated testing cycles. Both require deep domain expertise. Without automated drafting from data schemas and a testing environment tightly integrated with the data layer, this hands-on approach becomes slow and difficult to scale, delaying projects by weeks or months.

Knowledge graphs can theoretically bridge this gap. They capture not only what data exists, but what it represents and how it relates in a form that both humans and agents can understand. This would enable automatic drafting of tools and prompts from ontologies, while providing for an integrated testing environment and contextual reasoning with Agentic GraphRAG.

However, a concrete technical implementation has yet to be productized…until now.

Neo4j Aura Agent: Tight Knowledge Graph Integration Across the Agent Lifecycle

Neo4j Aura Agent solves these problems by integrating an agent-building platform directly into AuraDB, enabling users to build and deploy agents grounded in the context of knowledge graphs in minutes. Neo4j Aura Agent delivers four key advantages:

1. Graph-Driven AI – Auto-generate draft agents from your data schema and use case description(s)
2. Accurate Agentic GraphRAG – Robust retrieval customized to your knowledge graphs
3. Advanced Reasoning & Explainability – Transparent chain-of-thought multi-hop graph reasoning
4. Single-Click Deployment – Secure MCP and REST endpoints out-of-the-box

Let’s walk through how Neo4j Aura Agent addresses each stage of the agent lifecycle. You can also watch the demo to see each step in action: 

Graph-Driven AI: From Knowledge Graph to Agent in Minutes

Drafting first pass agents on top of enterprise data ordinarily requires experts who engineer prompts and tools by hand—a challenging process that can delay projects. But there’s enough information in a knowledge graph ontology to automate the initial agent construction.

Neo4j Aura Agent’s new “Create with AI” feature does exactly that. It takes two required inputs:

• A prompt from you describing the use case and what the agent will do
• The graph schema (pulled automatically from your selected graph database instance)

You can optionally provide vector embedding types for similarity search as well. 

Together, these constitute an ontology. Neo4j Aura Agent then automatically constructs an agent customized to your knowledge graph, complete with tailored prompts and graph retrieval tools—delivering a draft agent that’s ready to test and even deploy out-of-the-box in minutes, not weeks.

Accurate Agentic GraphRAG: Powerful, Customizable Retrieval

Once you have your initial agent, you can iteratively test and refine it using Neo4j Aura Agent’s low/no-code UI. Add, remove, or edit retrieval tools, adjust prompt instructions, and explore how different tools access your knowledge graph.

Specifically, Neo4j Aura Agent provides three primary types of graph retrieval tools:

• Similarity Search – Vector search for semantic retrieval
• Parameterized Query Templates – Pre-established graph patterns for critical, specialized queries
• Text-to-Query Generation – Dynamic Cypher query generation for flexible fallback

Your agent leverages these tools as needed to answer questions, delivering higher accuracy with more relevant, compact context than traditional RAG approaches. Knowledge graphs explicitly model the critical relationships in your data, providing the structure and relational context agents need to reason, respond, and act accurately.

Advanced Reasoning & Explainability: Trust Through Transparency

Neo4j Aura Agent leverages a full ReAct agent loop, enabling chain-of-thought reasoning. When combined with knowledge graph retrieval, this provides sophisticated multi-hop graph reasoning—the agent can traverse relationships in your data to pull together more complete, contextual information.

For transparency and trust, Neo4j Aura Agent exposes its reasoning to end users through a dedicated reasoning tab and structured response format. This explainability is critical for domains where accuracy is non-negotiable—pharma, legal, healthcare, financial services, and national security.

As Gartner puts it in a recent report: “True agentic AI systems require a context-aware data platform … featuring a rich semantic layer that enables AI agents to efficiently discover and safely access context-rich information.” That’s an excellent description of Neo4j Aura Agent and the Neo4j Graph Intelligence Platform.

Single-Click Deployment: Simplify Your AI Stack

When you’re ready to deploy, Neo4j Aura Agent makes it effortless. With a single click, you can deploy your agent to a secure, authenticated cloud endpoint with both a:

• REST API – Token-based authentication for programmatic access from any application
• Model Context Protocol (MCP) Server – OAuth-secured, cloud-hosted MCP server for maximum compatibility with AI clients including Claude, Cursor, Microsoft Copilot, ChatGPT, and more

No custom infrastructure setup. No weeks of LLM and embedding integration projects. Just production-ready deployment that frees you to focus on optimizing your agentic applications for your specific use cases.

Out-of-the-Box Production Infrastructure

Neo4j Aura Agent eliminates the complexity of agentic stacks with native end-to-end agent infrastructure powering everything from experimentation to testing and production deployment. Out of the box, you get:

• Google Gemini Flash 2.5 LLM for the agent runtime
• A specialized fine-tuned version of Gemini Flash for improved text-to-query retrieval
• Free enterprise-grade embeddings including Vertex AI gemini-embedding-001, text-embedding-005, text-multi-lingual-embedding-002, and Microsoft Azure OpenAI text-embedding-3-small/large/ada-002
• A simple agent UI for building, testing, debugging, and iterative prototyping

Realize Agentic AI’s Potential in the Real World

Enterprise AI leaders have been using Neo4j to ground agentic systems and improve their reasoning abilities. At QIAGEN and Daimler Truck North America, this has resulted in far more capable, versatile, and effective agentic applications. Neo4j Aura Agent now offers enterprises a rapid, dependable way to develop accurate and explainable agentic AI.

“Neo4j Aura Agent promises to improve healthcare by designing and deploying AI agents that create comprehensive knowledge graphs from our trusted biomedical knowledge. With new ways to interrogate these graphs, researchers can approach drug discovery in ways that were impossible before. That’s what makes this so promising for drug discovery and healthcare,” says Nitin Sood, Senior Vice President, Head of Product Portfolio and Innovation, opens in new tabQIAGEN.

“Enterprise knowledge graphs represent critical infrastructure for reliable agentic AI,” says Conor O’Shea, AI Architect, opens in new tabDaimler Truck. “At Daimler Truck North America, we’ve seen how Neo4j’s graph capabilities bring the accuracy and contextual reasoning that AI systems need to operate effectively in complex business environments. I’m excited to see how Neo4j Aura Agent and MCP Server will make these capabilities more accessible to enterprises building their next generation of intelligent applications.”

Refine and Evolve Your Agentic Applications

By simplifying agentic AI infrastructure and improving agent quality, Neo4j Aura Agent frees you to focus on what matters: optimizing your agentic offerings for specialized use cases. You can use the Neo4j Aura Agent UI to add or refine tools, redeploy agents in seconds, or spin up entirely new agents.

Neo4j Aura Agent is a natural fit for enterprise search and expert knowledge assistants. And it works equally well as a standalone API for building agentic systems or as a component within larger multi-agent architectures.

Get Started Today

For the month of February 2026, Neo4j Aura Agent is free to use. Starting in March, publicly accessible agents will be charged at a rate of $0.35 per agent hour. Neo4j Aura Agent is available across Free, Professional, and Business Critical Aura tiers.

Ready to build your first Neo4j Aura agent?

• Try the getting started tutorial
• Checkout the latest resources on the Neo4j Aura Agent webpage for additional learning, videos, and more.
• Start using Neo4j AuraDB for free.