Home › Claude & ChatGPT in DBeaver
SqlSage puts an AI SQL assistant inside DBeaver 26.x and signs in through the claude or codex CLI you already have. No API key, no separate AI bill, no token for SqlSage to keep. Usage counts against your existing Claude or ChatGPT plan.
Most AI tools ask you to paste a provider API key and then bill you per token on top of whatever you already pay. SqlSage does neither. It reuses the login that already lives in your claude or codex CLI — the same subscription you use in the terminal. SqlSage never asks for, stores or transmits an API key, and it never handles your provider tokens.
Concretely: to check whether you are signed in, SqlSage runs the CLI’s own status command as a short-lived process (claude auth status --json, or codex login status). That is the whole mechanism. Authentication stays inside the CLI where you control it; SqlSage only asks the CLI “is this session signed in?” and drives it to answer your questions.
Nothing to generate in a provider console, nothing to rotate, nothing that can leak from a preferences file.
AI usage counts against the Claude or ChatGPT/Codex plan you already have, under that provider’s terms. SqlSage adds no per-token AI fee.
SqlSage never sees, stores or transmits a credential. Sign-in lives in the CLI; the AI traffic is the CLI’s, not ours.
claude and use /login, or codex login. If you already use either in the terminal, you are done.Ctrl+Alt+A). The status bar shows which provider you are signed into.SELECT executes on a dedicated session and returns rows; anything that writes stops for your confirmation.PATH. Put the CLI’s directory on your user PATH and sign out and back in — full detail in the documentation.Cover Claude and ChatGPT/Codex from the same place: switch provider in the status bar and SqlSage re-checks sign-in for the one you picked. There is no separate “Claude edition” and “ChatGPT edition” — it is one plugin that speaks to whichever account you are logged into.
The model picker is editable: it suggests the aliases each CLI understands and accepts anything else you type, and blank means the CLI’s own default — so a model your account gains tomorrow needs no plugin update. The composed prompt goes to the CLI on standard input, never the command line, so it never shows up in a process list.
SqlSage always answers for one active connection, and the panel names which. Switch connections and the chat follows. The assistant grounds its SQL in the live schema — tables, columns, keys and indexes — instead of guessing identifiers.
The same keyless assistant works across the relational engines DBeaver connects to — PostgreSQL, MySQL/MariaDB, Oracle, SQL Server and SQLite — and writes SQL in each connection’s own dialect. If breadth across engines is what you care about, the companion page goes deeper: one copilot, five database engines →
Every statement the model proposes passes one classifier chokepoint before anything reaches your database. A plain SELECT runs; anything not proven to be a read — including writes, DDL and anything the parser cannot make sense of — is treated as a write and held for an explicit confirmation dialog showing the exact SQL. One statement per call: SELECT 1; DROP TABLE t is refused outright, never run.
You also get controls that sit outside the model’s reach: mark a connection Read-only (writes hard-refused) or No execution, and a global kill-switch that freezes every AI-initiated statement instantly. The honest limits — an engine-enforced read-only transaction only on PostgreSQL and MySQL/MariaDB, a query timeout only as reliable as the driver — are documented in full on the safety model page.
Only your SQL and schema metadata (table and column names and types) are sent to the model — never row data. When a read executes, the default is to return column names and the row count only, with values withheld. You can opt in to sharing result values with /share on or a Preferences checkbox; that opt-in is off by default and stays in effect until you turn it off.
Every classification decision and executed statement is appended to a local, hash-chained audit log on your own disk — nothing uploads it. And because SqlSage drives your local CLI, the AI traffic is the CLI’s conversation with Anthropic or OpenAI under your account, with credentials SqlSage never touches.
Sign-in, data and requirements. Engine and dialect questions are answered in the AI for DBeaver FAQ; everything else is in the documentation.
Yes — that is how it works. SqlSage starts the Claude Code (claude) or OpenAI Codex (codex) command-line tool you installed and signed in to on your own machine, and talks to it as a local process. Switch provider in the panel’s status bar and SqlSage re-checks sign-in for the one you picked. You need one of the two CLIs and a valid account with that provider for the AI features to work.
No. SqlSage never asks for, stores or transmits an API key, and it never handles your provider tokens. It reuses the login that already lives in your claude or codex CLI; to see whether you are signed in, it runs the CLI’s own status command (claude auth status --json or codex login status) as a short-lived process.
Sign in to the CLI once: run claude and use /login, or run codex login. If you already use either in the terminal, there is nothing more to do. If the CLI works in your terminal but the panel still reads “Not signed in”, DBeaver was most likely started from an icon and does not see your shell’s PATH: put the CLI’s directory on your user PATH, then sign out and back in. The step-by-step check is in the requirements guide.
Your SQL and schema metadata — table and column names, types, keys and indexes — go to the model. Row values do not: when a read runs, the model gets the column names and the row count, with values withheld. Sharing result values is an opt-in (/share on or a Preferences checkbox) that is off by default and stays on until you turn it off. That traffic is the CLI’s own conversation with Anthropic or OpenAI under your account; SqlSage’s tool channel is loopback-only, and the audit log and chat history stay on your disk.
Not for the AI. Usage counts against the Claude or ChatGPT/Codex plan you already have, under that provider’s terms, and SqlSage adds no per-token AI fee. The plugin itself has a Free tier and a 30-day Pro trial that starts on first run with no card and no sign-up; Pro plans are listed under pricing.
DBeaver 26.x, Community or PRO — SqlSage adds no DBeaver licence requirement. It needs Java 21, which DBeaver 26.x already bundles. The supported operating system is Windows; Linux and macOS are not supported and not tested. The reasons are in the operating-systems note.
Yes, through SqlSage’s outward MCP bridge, which is off by default. Tick Bridge for a connection and Copy config gives you a claude mcp add command (with snippets for Codex, Cursor and Claude Desktop too). Your agent then reaches that connection through the same chokepoint, over loopback only and with a bearer token — and the bridge forces Read-only with result values withheld, so writes are refused outright. Details: the bridge documentation.