SQLChat
A chat-based SQL client from Bytebase: ask questions in plain language and it queries your database.
About
SQLChat rebuilds the SQL client around a chat interface. Instead of navigating panels and writing queries by hand, you type questions in plain English and it generates the SQL, runs it against your database, and shows the results conversationally. It was built by Bytebase, the database DevOps company, as what they called their brick toward chat-era developer tools. It supports MySQL, PostgreSQL, SQL Server, TiDB Cloud, and OceanBase, and it can perform queries as well as modifications, additions, and deletions through the chat. The project is open source (MIT), built with Next.js, and ships as a Docker image you can self-host in one command.
On privacy, the project states that only your database schema is sent to the model API, never table data, and connection configs stay in your browser's local storage. One thing to know: the sqlchat.ai domain now redirects to bytebase.com, so the standalone hosted service appears to have been folded into Bytebase's main site; self-hosting via Docker remains the reliable way to run it. The honest trade-off is that generated SQL is still generated SQL: for destructive operations you should review the query before it runs, and schema-only context means the model can misread ambiguous column names.
What it can do
- Natural-language to SQL across five database engines
- Chat interface for querying, modifying, adding, and deleting data
- Schema-aware query generation (schema only sent to the model)
- Self-hosting via Docker; connection configs stored locally in the browser
- Multi-language UI (English, Chinese, Spanish)
Setup
- Self-host with Docker:
docker run -p 3000:3000 sqlchat/sqlchat. - Open the web UI and add your database connection (MySQL, PostgreSQL, MSSQL, TiDB Cloud, or OceanBase).
- Set your OpenAI API key in settings for the natural-language engine.
- Start chatting: ask questions in plain language and review the generated SQL before running destructive operations.
Prerequisites
Docker (for self-hosting) or use the hosted site; an OpenAI API key; database credentials.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: 5,842 GitHub stars (GitHub API), MIT license, pushed 2026-04-21, not archived; Bytebase's own blog confirms it built SQLChat; sqlchat.ai verified live but now redirects to bytebase.com. Strengths: clean scope, genuinely self-hostable, thoughtful privacy design. Docked for: (1) standalone hosted service folded into bytebase.com — self-host is the reliable path; (2) generated SQL still needs human review for destructive ops. License: MIT.
★ 5,842 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 9.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · 5,842 GitHub stars (GitHub API), pushed 2026-04-21, not archived (sqlchat/sqlchat)
- · Bytebase's own blog confirms it built SQLChat ('their brick toward chat-era developer tools')
- · sqlchat.ai verified live but now redirects to bytebase.com; self-host via Docker is the reliable path
- · README documents 5 supported DB engines, Docker quickstart, schema-only privacy model
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.