Natural Language SQL

Natural Language SQL turns questions written in plain English into database queries, checks and runs the generated SQL, then streams results to the user. It connects to PostgreSQL and MySQL, including multiple databases of the same type at once. A hybrid search using BM25 and sentence-transformer vectors selects schema details relevant to each question. Users can inspect and edit the SQL before execution and follow progress through query stages. Results can be sorted, paginated in 50-row pages, downloaded as CSV or JSON, or copied as tab-separated text. SELECT queries receive a default limit; explicit result counts are capped at 1,000 rows, and execution has a 30-second default timeout. INSERT, UPDATE, and DELETE are enabled by default with warning banners. Ollama runs locally by default, keeping questions, schema, and generated SQL on the user’s machine; OpenAI, Gemini, and Groq are optional cloud providers. The project is free under GPL-3.0. Its recommended Docker setup calls for Docker 20.10+ with Compose v2 and about 10 GB of free disk space; the first run downloads an approximately 5 GB model.

Natural Language SQL combines plain-language querying with editable SQL and practical result handling for PostgreSQL and MySQL. Its default local Ollama option keeps query context on the machine, while the recommended Docker setup has substantial disk and first-run download requirements.

It suits people who want to ask questions of PostgreSQL or MySQL databases in plain English while retaining the option to review SQL before it runs. Those preferring the default local AI setup can keep questions, schema, and generated SQL on their machine.

Overview

Natural Language SQL is an open-source tool for turning database questions into SQL and returning query results. It is a strong fit for developers who want a self-hosted workflow with editable queries; the local-AI option is appealing, though the recommended setup needs considerable disk space and writes are enabled by default.

Key features

Hybrid retrieval combines BM25 search with sentence-transformer vectors to find schema context relevant to a question. That can reduce the need to assemble context manually, but generated SQL still deserves scrutiny: the project reports 52.32% execution accuracy on Spider dev.

Users can review and edit SQL before execution, with progress streamed through the query stages. Explanations, sorting, pagination in 50-row pages, CSV and JSON downloads, and tab-separated copying make results easier to inspect and pass along. SELECT queries receive a default LIMIT; explicit result counts are capped at 1,000 rows, and execution times out after 30 seconds by default.

INSERT, UPDATE, and DELETE are allowed by default, with warning banners. Those warnings do not block changes, so the project recommends connecting with a read-only database user when writes should be prevented.

Ollama runs locally by default: questions, schema, and generated SQL stay on the user's machine, and the project says it collects no telemetry. With OpenAI, Gemini, or Groq, the question, relevant schema, and SQL context go to that provider; database passwords and result rows are excluded. Passwords and cloud API keys are encrypted at rest, redacted from logs, and masked in API responses. Session query results, schema metadata, and history remain in memory rather than being written to disk by the application.

Pricing

The Natural Language SQL Engine costs nothing under the GPL-3.0 open-source license, for local installation. There is no paid plan to unlock in the supplied pricing structure, making it a fit for users willing to manage their own deployment rather than pay for a hosted service. The trade-off is setup effort: the recommended Docker deployment needs Docker 20.10 or later, Compose v2, about 10 GB of free disk space, and an initial download of an approximately 5 GB model.

Platforms

Natural Language SQL is available as an API, on Linux and macOS, as self-hosted software, and on the web. It connects to PostgreSQL and MySQL and can query multiple databases of the same type in parallel.

Who it's for

Choose it if you work with PostgreSQL or MySQL, want to ask questions in plain language, and value seeing and editing SQL before it runs. Its local default suits users who want query context to stay on their machine. It is less suitable for anyone who needs writes blocked by default, wants to avoid managing a local deployment, or cannot spare the disk space for the recommended Docker setup.

Pros and cons

• Pros: Local Ollama keeps questions, schema, and generated SQL on the machine; SQL is editable before execution; results can be sorted, paginated, downloaded, or copied; the engine is GPL-3.0 open source.
• Cons: The recommended Docker setup needs about 10 GB of free disk space and downloads an approximately 5 GB model; write queries are allowed by default; cloud providers receive question, schema, and SQL context.

Alternatives

AI SQL Generators is the broader directory to consult when comparing tools in this category.

Pick Outerbase AI if you need a free plan for up to five users and want a hosted-style tier with 10 EZQL queries per month, one base, three saved queries, and one dashboard; its free plan is limited to transactional databases.

SQL Mocker is an option for web users who want a free tier with 50 AI queries total and five saved projects, or a one-user Starter tier with 100 AI queries per month for $10 a month.

Consider Wren AI if an open-source context engine used through CLI and MCP without a UI better suits your workflow.

NatureQuery may suit web and API users who prefer a free hosted allowance of 50 queries per month, one database connection, exports, and 30-day query history.

QueryPlane offers a self-hosted and web option with a free tier capped at three apps, one database connection, and 50 AI prompts per month.

SQLer is a web and API alternative with 50 NL2SQL queries per month on its free plan.

IDERA AI SQL Query Generator is another free web-based option.

mnemiq is worth considering if you want an open-source Apache-2.0 engine in your own environment.

Verdict

Natural Language SQL is a good choice for developers who want an open-source, self-hosted natural-language interface to PostgreSQL or MySQL, especially when keeping AI context local matters. Its editable SQL and practical result handling strengthen the case, but the sizeable local setup and default write access make it a poor fit for users seeking a low-maintenance or write-locked workflow.