Postquel: a native PostgreSQL client for Mac, with your own AI

Описание и применение

Postquel — нативный клиент PostgreSQL для macOS, интегрирующий AI без отдельного API‑ключа, позволяющий просматривать и редактировать таблицы, генерировать SQL из естественного языка и безопасно управлять схемой.

Задача

Database developers need a lightweight, native macOS client that integrates AI assistance without requiring separate API keys.

Результат

Postquel enables developers to browse and edit PostgreSQL databases natively on macOS while leveraging existing AI accounts for read‑only data queries and SQL generation, eliminating the need for separate API keys and simplifying workflow.

Рабочий процесс

• Launch Postquel and connect to a PostgreSQL database via libpq.
• Browse tables in the sidebar and open a table to view Content, Structure, or DDL.
• Edit rows directly in the grid or follow foreign keys to related tables.
• Use Cmd+Return to run a single query in a file, or save queries to the sidebar.
• Edit schema changes (columns, indexes, constraints) and preview generated SQL before applying in a single transaction.
• Open the assistant panel and sign in to an existing AI service (Claude Code, Codex, or Cursor Agent).
• Ask natural language questions about data (e.g., "Which customers spent the most last month?") and receive SQL queries and results in the chat window.
• Use the prompt bar to convert natural language descriptions into SQL queries.
• Review any proposed schema or data changes suggested by the assistant before executing them.
• Use SSH tunnels for remote connections and store passwords in Keychain.
• Run the app on macOS 15+ on Apple silicon or Intel; no local Postgres installation required.

Доказательства

If you're signed in to Claude Code, Codex or Cursor Agent, Postquel can use that. You pick the agent and model. There's no separate API key to put into the app.

Источник

The assistant has tools for reading the schema and querying your connected database. It can look up the tables itself. You don't have to copy a bunch of table definitions into a chat window first.

Источник

The assistant's database connection is read-only. If it wants to change data or the schema, it proposes SQL for you to review, dry-run and apply.

Источник