Describe what you want in plain language and get real SQL against your real schema: SQL Chat is an open-source, chat-based SQL client from the Bytebase team. Instead of writing queries in a traditional editor, you connect a database and describe what you want in plain language; the AI reads your schema automatically, generates SQL that references real table and column names, executes it, and returns tabular results in the conversation. Follow-up messages refine the query, so exploration becomes a dialogue - narrow a result set, add a join, change an aggregation - without retyping statements. It supports MySQL, PostgreSQL, SQL Server, TiDB Cloud, and OceanBase from one interface, and covers modification as well as reads: insert, update, and delete operations phrased conversationally. Built with Next.js and TypeScript, it deploys as a single stateless Docker container in single-user mode - connection profiles live in the browser, so there is nothing server-side to maintain. A custom AI endpoint setting routes inference through any OpenAI-compatible API, including self-hosted models, and an optional database-backed mode adds accounts and quotas for offering the tool to a team. MIT-licensed.

For developers and database administrators who want a full database IDE inside the browser without installing separate desktop clients on every machine, OrcaQ provides a unified workspace for PostgreSQL, MySQL, MariaDB, Oracle, SQLite, Cloudflare D1, Turso, and Redis. Engineers organize connections into named workspaces with SSH bastion tunneling, granular SSL configuration, and color-coded environment tags whose strict mode requires explicit confirmation before executing destructive mutations against production databases. The Quick Query data grid lets users filter, sort, paginate, and edit table rows inline while jumping across foreign-key relationships and inspecting table DDL metadata. When writing complex multi-statement SQL, the Raw SQL Editor provides context-aware autocomplete, query variables, nested script folders, and visual EXPLAIN execution plan analysis. Database architects can render interactive entity-relationship diagrams for single tables or entire schemas, compare two database schemas side by side to generate safe migration scripts, and manage Redis keys across string, hash, list, set, and sorted-set structures in a dedicated workbench. PostgreSQL administrators also gain a visual role-permission matrix, real-time instance health monitoring for active sessions, locks, and replication lag, native backup and restore job tracking, and an optional AI Orca Agent that drafts queries and detects anomalies while requiring human approval before any data modification runs. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.