Last Updated: 8 September 2026
RavenDB has launched Quill, a new product that connects AI agents to information stored in existing SQL databases. Companies can use the system without moving their data to a new database or replacing the system they already rely on.
Many companies keep important operational data in SQL databases that were built before technologies such as AI agents and embeddings became common. Connecting that information to AI applications can require new data pipelines, search tools, security controls and other infrastructure. That work can take considerable time.
RavenDB describes Quill as a context layer placed between a company's database and its AI applications. In simple terms, the layer prepares selected information so an AI agent can find and use it while leaving the original SQL database as the authoritative source. The existing database stays in place.
Quill currently works with PostgreSQL, Microsoft SQL Server and MySQL. Organizations can deploy it in the cloud or on their own infrastructure, depending on their technical, regulatory or data residency requirements.
Once connected to a database, Quill keeps track of changes in the source information and updates the data available to AI applications. RavenDB says the system includes embeddings, vector indexes and full text search. Embeddings turn information into numerical representations that help AI systems identify related content, while vector indexes make that information easier to retrieve.
RavenDB says using the product can reduce the time required to launch a production AI agent from the 18 to 24 months associated with a typical internal build to a matter of weeks. Quill also includes search, retrieval and agent infrastructure.
"Anyone can stand up an AI demo in an afternoon, but getting that demo into production with data pipelines, semantic search, security, governance, all the plumbing a small proof of concept doesn't need until it has to run at scale, is the hard part. Quill exists because we'd rather hand teams that plumbing already assembled than watch them rebuild the same project after project. You get access to the live data you need, decide the scope on day one, and change it as you go, instead of building everything from scratch."
Oren Eini, founder and CEO of RavenDB
Quill provides controls over which information an AI agent is allowed to access. The system sits between the AI application and the source database, giving organizations another place to define what data can be used.
According to RavenDB, Quill is designed on the assumption that an AI model should not receive unrestricted access to a company's information. Organizations can decide which data enters the context layer and determine what individual agents are permitted to see. Those permissions can differ between agents.
RavenDB gives healthcare as one example. An organization could allow an agent to retrieve details about a patient's upcoming appointment while keeping prescription information outside the dataset available to that agent. These controls operate independently from the permissions already configured in the original database.
Quill is model agnostic, which means it is not tied to a particular AI model or provider. Companies can choose the models they want to use, change providers later or run models on their own hardware.
The product includes search, retrieval and agent components for applications that use information from the connected database. RavenDB says agents created with Quill can be made available through web chat, WhatsApp, Telegram, Slack and Discord.
Albos Technologies and Holdings is an early user of Quill.
"With Quill, the plumbing was already there, so we spent our time building the actual feature."
Hagay Albo, CEO of Albos Technologies and Holdings
Quill follows several AI related additions to RavenDB's technology. The company's product roadmap shows that it introduced vector search and embedding generation in early 2025, followed by generative AI capabilities and AI agents later that year. In 2026, RavenDB added features including an AI assistant, AI agent attachments and support for multiple agents.
RavenDB was originally developed as a NoSQL document database. Unlike a traditional relational database, which commonly organizes information into tables and rows, a document database stores information in flexible documents that can contain different types of data.
RavenDB 1.0 was released in 2010 for the .NET software ecosystem and included ACID transaction support, a set of database properties designed to keep transactions reliable. The company says its database technology is now used by more than 12,000 customers across 50 industries.
Its existing database product includes automatic indexing, full text search and tools for monitoring and administration.
Quill allows RavenDB to work with organizations that continue to use SQL as their main database. Those companies do not need to replace their SQL system with RavenDB as the system of record. Instead, Quill operates above the existing database and provides selected information to AI applications.
Quill is available for organizations using PostgreSQL, SQL Server or MySQL. RavenDB says support for additional databases is planned, and the product can be deployed in the cloud or on premises.
