Dbhub is live · free · last verified 25 Oct 2026 · viability monitor
Dbhub
A minimal, MIT-licensed database MCP server that connects Claude Code, Cursor, and VS Code to Postgres, MySQL, SQL Server, MariaDB, Oracle, and SQLite.
If you point MCP clients at more than one database engine, DBHub is the least wasteful way we've seen to do it: one process, ~1.4k tokens of tool overhead versus MCP Toolbox's 19.0k and Supabase MCP's 19.3k, and read-only mode plus row limits and query timeouts that make production diagnostics survivable. Set expectations — configuration is TOML and the command line, and the Workbench shows request traces, not a query grid. Use it to give agents schema-aware SQL access; keep DBeaver or TablePlus for hands-on SQL work.
Verdict generatedVerified 5d ago · liveness 67/100 · cite: rightaichoice.com/tools/dbhub
- Developers running Claude Code, Cursor, or VS Code who want database access without burning context on tool definitions
- Teams with a mixed database fleet replacing several single-engine MCP servers with one DBHub process
- Production troubleshooting where read-only mode, row limits, and query timeouts keep diagnostics contained
- Teams that prefer Docker/NPM installs and MIT-licensed, zero-dependency tooling
- Anyone wanting a full GUI database client with sortable grids and visual query builders
- Teams needing change data capture or streaming changes out of a database
- Users new to MCP or the command line who want a graphical setup wizard
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip DBHub if you want a visual query grid and a point-and-click connection wizard, or if your workflow depends on a broad catalog of prebuilt tools rather than two core ones plus custom SQL.
DBHub is MIT-licensed and installs from Docker, NPM, the MCP Bundle, or the Claude Code plugin, so the cost of adopting it is engineering time on a TOML config rather than a subscription. The real comparison is overhead: ~1.4k tokens of tool definitions versus 19.0k for MCP Toolbox's 28 defaults and 19.3k for Supabase MCP's full set.
In short
Dbhub — A minimal, MIT-licensed database MCP server that connects Claude Code, Cursor, and VS Code to Postgres, MySQL, SQL Server, MariaDB, Oracle, and SQLite. Best for Developers running Claude Code, Cursor, or VS Code who want database access without burning context on tool definitions, Teams with a mixed database fleet replacing several single-engine MCP servers with one DBHub process, Production troubleshooting where read-only mode, row limits, and query timeouts keep diagnostics contained. Free to use.
What is Dbhub?
Dbhub — A minimal, MIT-licensed database MCP server that connects Claude Code, Cursor, and VS Code to Postgres, MySQL, SQL Server, MariaDB, Oracle, and SQLite. Best for Developers running Claude Code, Cursor, or VS Code who want database access without burning context on tool definitions, Teams with a mixed database fleet replacing several single-engine MCP servers with one DBHub process, Production troubleshooting where read-only mode, row limits, and query timeouts keep diagnostics contained. Free to use.
DBHub is a minimal database MCP server from Bytebase that acts as one gateway between MCP-compatible AI clients and PostgreSQL, MySQL, SQL Server, MariaDB, Oracle, and SQLite. You install it via Docker, NPM, the MCP Bundle, or the Claude Code plugin, declare your connections in TOML, and then point Claude Desktop, Claude Code, Cursor, or VS Code at it. It is the official example in the Claude Code docs for connecting to PostgreSQL over MCP. The design bet is context economy. DBHub loads two tools by default — execute_sql and search_objects — at roughly 1.4k tokens, against 19.0k for MCP Toolbox's 28 default tools and 19.3k for Supabase MCP's full set, per the vendor's own comparison table. execute_sql runs queries with transaction support and safety controls; search_objects walks schemas, tables, columns, indexes, and procedures so the model can orient itself before writing SQL. Two more tools are opt-in: explain_sql returns an execution plan without running the query, and health_check reports connection pool state plus buffer cache hit ratio. Custom Tools let you publish reusable parameterized SQL operations. Guardrails are what make it usable against live databases: read-only mode, row limiting, and query timeouts to contain runaway operations, plus SSH tunneling and SSL/TLS for the connection itself. Multi-connection TOML config lets one DBHub process front several databases, and a browser-based Workbench runs tools and shows request traces when you don't want an MCP client in the loop. Version 1.0.0 (2026-07-28) adopts the MCP 2026-07-28 spec with a stateless model, cacheable tool lists, and header-based routing, and stays backwards compatible with 2025-era clients.
Key features
- Minimal database MCP server loading 2 default tools at ~1.4k tokens
- execute_sql tool for running queries with transaction support and safety controls
- search_objects tool for exploring schemas, tables, columns, indexes, and procedures
- Opt-in explain_sql tool returning a query execution plan without executing it
- Opt-in health_check tool reporting connection pool state and buffer cache hit ratio
- Custom Tools for defining reusable, parameterized SQL operations
- Supports PostgreSQL, MySQL, SQL Server, MariaDB, Oracle, and SQLite
- Multi-connection TOML configuration for several databases in one process
- Read-only mode, row limiting, and query timeouts as query guardrails
- SSH tunneling and SSL/TLS encryption for database connections
- Browser-based Workbench for running tools and viewing request traces
- MCP 2026-07-28 spec support in v1.0.0 with stateless model and cacheable tool lists
- Header-based routing under the MCP 2026-07-28 spec
- Backwards compatible with 2025-era MCP clients
- Zero-dependency install via Docker, NPM, MCP Bundle, or Claude Code plugin
Behind the verdict
DBHub's whole thesis is that most database MCP servers waste your context window. Its own comparison puts it at 1.4k tokens for 2 default tools, against 19.0k for MCP Toolbox's 28 defaults and 19.3k for Supabase MCP's full set — a 13-14x difference that matters when that context is competing with your actual code. Whether or not you agree with the exact numbers, the architectural choice is real: execute_sql and search_objects by default, with explain_sql and health_check opt-in, so you pay for tools you actually call. Where it earns its place is the fleet case. One DBHub process, a TOML file with multiple connections, and PostgreSQL, MySQL, SQL Server, MariaDB, Oracle, and SQLite all reachable through a single interface — that replaces several single-engine MCP servers and the per-server token tax that comes with them. The guardrails are the other half: read-only mode, row limiting, and query timeouts stop a confident model from running away on a production database, and SSH tunneling plus SSL/TLS cover the connection leg. search_objects walking schemas, tables, columns, indexes, and procedures means the model can look before it writes SQL rather than guessing column names. Custom Tools are the underrated feature. You define a reusable, parameterized SQL operation once and hand it to a non-technical colleague through Claude Desktop as a curated read-only view, instead of exposing raw SQL. The Workbench then lets you run those tools in a browser and see the request trace when you want to verify what the model actually did. The trade-offs are real. The tool surface is deliberately small, so if your workflow depends on a wide catalog of prebuilt operations, MCP Toolbox's 28 tools are the opposite bet. Everything is configured in TOML and driven from the command line — no graphical setup wizard — and the Workbench is a tool runner and request tracer, not a database client with sortable grids or a visual query builder. Nor does DBHub do change data capture or stream changes out of a database. Version 1.0.0 (2026-07-28) moves to the MCP 2026-07-28 spec with a stateless model, cacheable tool lists, and header-based routing, while remaining backwards compatible with 2025-era clients — worth knowing if your MCP client fleet is mixed. It is MIT-licensed and installs from Docker, NPM, the MCP Bundle, or the Claude Code plugin with no dependency chain to manage.
Limitations
- DBHub is a minimal MCP server with a small tool surface — 2 default tools (execute_sql, search_objects) at ~1.4k tokens, with explain_sql, health_check, and Custom Tools as opt-ins.
- It reaches PostgreSQL, MySQL, SQL Server, MariaDB, Oracle, and SQLite through multi-connection TOML configuration and installs zero-dependency via Docker, NPM, MCP Bundle, or the Claude Code plugin.
- Guardrails are read-only enforcement, row limiting, and query timeouts.
- The Workbench is a tool runner and request tracer, not a database client, and configuration is TOML plus the command line rather than a graphical wizard.
- There is no change data capture or change streaming.
as of 2026-10-20
Researching Dbhub? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Dbhub still active?
Dbhub is live · free · last verified 25 Oct 2026 · viability monitor.
Viability score
How well maintained and how widely used is Dbhub? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What's new in Dbhub
Checked 10 days agoAcross the latest 1 update: 1 launch.
What do users say about Dbhub?
What people actually say about Dbhub — is it worth it?
Our community sentiment research written Jul 30, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
How we verified this
Every listed fact about Dbhub— pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 25 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified Dbhub 7 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed description, limitations, our verdict, who it suits, setup time, features, use cases, pricing fit, hidden costs, who should skip it, switching out, switching in, summary
- — changed our verdict, features, who should skip it, description, who it suits, summary
- — changed pricing page
- — changed use cases, description, our verdict, features, hidden costs, limitations, switching in, switching out, who should skip it, pricing fit, setup time, summary
- — changed SSL certificate
- — changed limitations
Showing the 6 most recent of 7 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Dbhub cost in 2026?
Dbhub: Free to use. Pricing last verified 25 Oct 2026.
Where the pricing makes sense
The company stage and team size where Dbhub's pricing actually pencils out — and where peers do it cheaper.
DBHub is MIT-licensed and installs from Docker, NPM, the MCP Bundle, or the Claude Code plugin, so the cost of adopting it is engineering time on a TOML config rather than a subscription. The real comparison is overhead: ~1.4k tokens of tool definitions versus 19.0k for MCP Toolbox's 28 defaults and 19.3k for Supabase MCP's full set.
Has Dbhub's pricing changed?
No plan price for Dbhub has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Dbhub?
Pick Dbhub if you are Developers running Claude Code, Cursor, or VS Code who want database access without burning context on tool definitions, Teams with a mixed database fleet replacing several single-engine MCP servers with one DBHub process, Production troubleshooting where read-only mode, row limits, and query timeouts keep diagnostics contained, Teams that prefer Docker/NPM installs and MIT-licensed, zero-dependency tooling.
Look elsewhere if you are Anyone wanting a full GUI database client with sortable grids and visual query builders, Teams needing change data capture or streaming changes out of a database, Users new to MCP or the command line who want a graphical setup wizard, Workflows that depend on a large default tool catalog like MCP Toolbox's 28 tools.
Skip it if Skip DBHub if you want a visual query grid and a point-and-click connection wizard, or if your workflow depends on a broad catalog of prebuilt tools rather than two core ones plus custom SQL.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Dbhub actually fits — and what changes day-one when you adopt it.
You add DBHub to Claude Code, declare the staging Postgres connection in TOML, and keep read-only mode on while the agent calls search_objects to learn the schema and execute_sql to check data.
Outcome: The agent writes SQL against real column names instead of guessing, and the whole server costs ~1.4k tokens of context rather than 19k.
You replace separate single-engine MCP servers with one DBHub process whose TOML declares connections to PostgreSQL, MySQL, and SQL Server, then point Claude Desktop, Cursor, and VS Code at the same /mcp endpoint.
Outcome: One process and one config file front the whole fleet, and every client shares the same tool definitions.
You define Custom Tools as parameterized read-only SQL operations, enable read-only mode and row limiting, and hand colleagues Claude Desktop pointed at DBHub.
Outcome: Colleagues ask questions and get answers from parameterized operations you wrote, and you can inspect each request in the Workbench request trace.
Use cases
- Query databases directly from Claude Desktop using natural language
- Explore a PostgreSQL schema and run SQL through the VS Code MCP integration
- Build AI agents that read and write across several database engines with guardrails
- Run read-only diagnostics against production with row limits and query timeouts
- Give non-technical staff curated, read-only database views via Claude Desktop
- Prototype SQL operations once and reuse them as parameterized Custom Tools
- Consolidate separate per-database MCP servers into a single DBHub process
- Set up local development with read-only SQLite access for AI assistants
Setup time & first value
How long it actually takes to get something useful out of Dbhub — broken out by persona, not the marketing-page minute.
Rough ETA: an experienced developer gets a first working connection in well under 30 minutes — install via Docker, NPM, MCP Bundle, or the Claude Code plugin, write the TOML connection block, point the client at the /mcp endpoint. Budget longer for SSH tunneling, SSL/TLS, and multiple connections.
Switching to or from Dbhub
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From per-engine MCP servers: declare each database as a connection in one DBHub TOML file and retire the individual servers.
- →From MCP Toolbox: keep your SQL as Custom Tools, since DBHub ships only execute_sql and search_objects by default.
- ↗To MCP Toolbox: port Custom Tools into its tool catalog if you need a large default tool set rather than two core tools.
- ↗To DBeaver or TablePlus: keep DBHub for agent access and add a GUI client for hands-on grid work, since the Workbench is a tool runner, not a query grid.
Integrations
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Dbhub”, and we withheld 6: 6 could not be judged, because “Dbhub” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Dbhub.
Official links
What are the best Dbhub alternatives?
The closest alternatives to Dbhub we list are Mcp Toolbox, TablePlus and Context7. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Mcp Toolbox | Free | Free | — |
| TablePlus | Paid | Pricing not published | — |
| Context7 | Freemium | Pricing not published | — |
Alternatives to Dbhub
View all
Mcp Toolbox
Open-source MCP server for natural-language SQL and Google Cloud admin operations across 40+ data sources
Tools that pair well with Dbhub
Common stack mates teams adopt alongside Dbhub, with the specific reason each pairing earns its keep.
Mcp Toolbox
Open-source MCP server for natural-language SQL and Google Cloud admin operations across 40+ data sources
TablePlus
Native database GUI for MySQL, PostgreSQL, SQLite and more, bought once instead of rented
Context7
Context7 delivers version-pinned library documentation into Claude, Codex and Cursor agents so they stop inventing APIs.
Dbhub vs its closest rival
Dbhub's closest head-to-head is with Spider Cloud — These are not substitutes — they are two different organs in the same agent stack. Spider Cloud reaches outward (rendering, crawling, search, proxies, CAPTCHA handling) and bills per use; DBHub reaches inward (your own SQL databases) and costs nothing but a TOML file. If your agent needs live pages, buy Spider. If it needs to query a database, install DBHub. The only overlap is that both can be wired into Claude Code or Cursor, and self-hosting DBHub removes even that overlap. A buyer shortlisting one for the other's job has misread the problem. Verdict generated
Head-to-head comparisons
Dbhub vs Spider Cloud
These are not substitutes — they are two different organs in the same agent stack. Spider Cloud reaches outward (rendering, crawling, search, proxies, CAPTCHA handling) and bills per use; DBHub reaches inward (your own SQL databases) and costs nothing but a TOML file. If your agent needs live pages, buy Spider. If it needs to query a database, install DBHub. The only overlap is that both can be wired into Claude Code or Cursor, and self-hosting DBHub removes even that overlap. A buyer shortlisting one for the other's job has misread the problem.
GeneratedDbhub vs Dbos
Choose DBOS if you need fault-tolerant, durable execution for AI agents or business workflows and already use Postgres. Choose DBHub if you want a lightweight, token-efficient MCP server to give AI coding assistants (Claude, Cursor, etc.) direct, secure access to multiple database types. They solve different problems: DBOS is for orchestrating complex, stateful processes; DBHub is for database querying from AI tools.
GeneratedDbhub vs Temporal Ai
Temporal AI and Dbhub solve completely different problems. Temporal is a heavy-duty durable execution platform for building reliable AI agents and long-running workflows — think OpenAI and Lovable. Dbhub is a lightweight, token-efficient MCP server that gives AI assistants direct SQL access to databases. If you need crash-proof AI orchestration, choose Temporal; if you want Claude to query your local PostgreSQL, choose Dbhub.
GeneratedDbhub vs Voyage Ai
These are not competitors and nobody is choosing between them. Voyage AI is an enterprise retrieval-model purchase — you buy embeddings, rerankers, multimodal support, and 32K-token document context to raise RAG answer accuracy, and you pay through a sales conversation. DBHub is free infrastructure that lets Claude Code, Cursor, or VS Code query your existing PostgreSQL/MySQL/SQL Server/MariaDB/SQLite databases while loading only ~1.4k tokens of tools. The realistic scenario is stacking them: DBHub pulls rows and schema out of operational databases, Voyage embeds and ranks what you index. If your problem is retrieval quality on finance or legal corpora, DBHub is irrelevant. If your problem is giving an agent SQL access without blowing its context window, Voyage is irrelevant.
GeneratedDbhub vs Bito
If your team struggles with cross-repo dependencies and needs architectural context for AI coding agents, Bito is the obvious choice despite its opaque pricing. For developers who just want a lightweight, open-source MCP gateway to databases, DbHub is a perfect free tool. They solve entirely different problems—choose based on whether you need system-wide context or database connectivity.
GeneratedFrequently asked questions
Used Dbhub? Help shape our editorial sentiment research.