QueryWeaver: Ask Your Database Questions in Plain English

By Prahlad Menon 5 min read

If you’ve ever stared at a database schema trying to figure out which tables to join, QueryWeaver might save you a lot of headaches. It’s an open-source tool that lets you ask questions about your data in plain English and get back working Structured Query Language (SQL) queries. Built by FalkorDB, a graph database company, it takes a unique approach to the Text2SQL problem — and it’s worth a closer look.

What Is QueryWeaver?

QueryWeaver is an open-source Text2SQL tool that converts natural language questions into SQL queries. Instead of memorizing table names, column types, and join conditions, you type something like “show me the top 10 customers by revenue last quarter” and QueryWeaver generates the SQL for you.

What sets it apart from other Text2SQL tools is how it understands your database. QueryWeaver maps your schema as a knowledge graph using FalkorDB, so the Large Language Model (LLM) powering the translation actually understands the relationships between your tables — not just their names.

How Does QueryWeaver Convert Natural Language to SQL?

Most Text2SQL tools dump your schema into an LLM prompt as flat text. QueryWeaver does something smarter: it builds a graph representation of your database schema, capturing foreign keys, relationships, and table hierarchies as connected nodes.

When you ask a question, QueryWeaver traverses this graph to pull in only the relevant schema context. This means the LLM gets a focused, relationship-aware view of your data instead of a wall of CREATE TABLE statements. The result is more accurate queries, especially on complex schemas with dozens of tables.

What Databases Does QueryWeaver Support?

QueryWeaver works with any database you can connect to via standard drivers. The graph-powered schema understanding is database-agnostic — it reads your schema metadata and builds the knowledge graph regardless of whether you’re running PostgreSQL, MySQL, or another relational database.

How Do You Set Up QueryWeaver?

Getting started is a single Docker command:

docker run -p 5000:5000 -it falkordb/queryweaver

That spins up the QueryWeaver server on port 5000. From there, you connect your database, and QueryWeaver automatically maps your schema into its graph representation. There’s also a hosted version at app.queryweaver.ai if you’d rather skip the self-hosting.

QueryWeaver supports multiple LLM backends — OpenAI, Google Gemini, Anthropic, and Azure — so you can use whichever provider fits your setup and budget.

Does QueryWeaver Have an API?

Yes, and this is where it gets interesting for developers. QueryWeaver exposes a Representational State Transfer (REST) Application Programming Interface (API) for programmatic access. You can send natural language queries via HTTP and get SQL back, making it easy to integrate into dashboards, internal tools, or automation pipelines.

Full API documentation is available at app.queryweaver.ai/docs.

What Is QueryWeaver’s MCP Server?

QueryWeaver also ships with a Model Context Protocol (MCP) server. MCP is an emerging standard that lets AI agents interact with external tools natively. If you’re building AI agents with frameworks like LangChain or using Claude’s tool-use capabilities, the MCP server means your agent can query databases through QueryWeaver without any custom integration code.

This is a big deal for the agentic AI space — your AI assistant can answer data questions by generating and running SQL on the fly.

Does QueryWeaver Remember Previous Conversations?

Yes. QueryWeaver stores per-user conversation memory in FalkorDB with a configurable Time To Live (TTL) via the MEMORY_TTL_SECONDS setting. This means follow-up questions work naturally — you can ask “now filter that by region” and QueryWeaver remembers what “that” refers to.

This conversational context makes it feel less like a query generator and more like a data analyst you’re chatting with.

Is QueryWeaver Secure?

QueryWeaver includes OAuth support for both Google and GitHub, so you can control who has access. Combined with the self-hosting option, teams can keep everything behind their own firewall while still getting the convenience of natural language database access.

What This Means For You

For developers: QueryWeaver’s REST API and MCP server mean you can add natural language database access to any application with minimal effort. The Docker deployment keeps infrastructure simple.

For data analysts: Instead of writing complex joins from scratch or waiting for engineering to build a report, you can explore data conversationally. The graph-powered schema understanding handles the relationship logic that usually trips people up.

For teams: Non-technical stakeholders can ask data questions directly. The conversation memory means they can iterate on queries without starting over each time. OAuth keeps access controlled.

For AI builders: The MCP server makes QueryWeaver a natural fit for agentic workflows. If you’re building AI systems that need to answer questions from structured data, this is a ready-made tool rather than something you’d need to build yourself.

Is QueryWeaver Really Free?

QueryWeaver is fully open source and available on GitHub. You’ll need your own LLM API keys (OpenAI, Gemini, Anthropic, or Azure), but the tool itself is free to use and self-host. The hosted version at app.queryweaver.ai is available if you prefer a managed experience.