20 Best PostgreSQL MCP Servers, Compared (October 2026)
The short answer
Most developers who want to connect an agent to PostgreSQL this afternoon should start with DBHub (bytebase/dbhub): it is a Community favourite, has 3,639 GitHub stars, gained 125 stars per 30 days, had a commit 5 days ago, has 94 commits on its default branch in the last 12 weeks, and executes SQL under read-only mode, row limiting, and query timeouts. If your priority is PostgreSQL performance tuning and health checks rather than a general database client, Postgres MCP (crystaldba/postgres-mcp) is the better answer: it is a Community favourite with 3,402 GitHub stars and 89 stars gained per 30 days, its tool descriptions are graded A, and it provides tools for index recommendations, EXPLAIN plans, top queries, database health, and safe SQL execution. DBHub is the more actively committed choice, with 94 commits on its default branch in the last 12 weeks versus 1 for Postgres MCP (crystaldba/postgres-mcp).
Whichever you choose, give it the narrowest access that still works (a read-only credential, a replica, a scratch account), and widen it only once you have watched what your agent actually asks for.
Glama operates the MCP registry these numbers are measured from, and sells MCP hosting and a gateway. No position on this page is paid for. How the registry is built.
Quick picks
- 1pg-aiguide : AI coding assistants writing PostgreSQL code: it provides semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation plus installable Postgres Agent Skills.
- 2Postgres MCP : PostgreSQL performance tuning and health checks: it exposes 9 tools for index recommendations, EXPLAIN plans, top queries, database health, and safe SQL execution.
- 3pgEdge Postgres MCP Server : Internal developer workflows against trusted PostgreSQL databases: it provides SQL query access through MCP-compatible clients and a natural language agent for forming SQL queries.
- 4DBHub : Local development and read-only production troubleshooting alongside coding agents: it executes SQL under read-only mode, row limiting, and query timeouts.
- 5MCP PostgreSQL Operations : Monitoring PostgreSQL 12-18 in production with read-only analysis: it exposes 34 tools for locks, WAL, replication, bloat, vacuum, query, and schema insights.
Which one, for your situation
| Your situation | What to use |
|---|---|
| I need a universal database client with read-only safeguards | Use DBHub (bytebase/dbhub): it is a Community favourite, had a commit 5 days ago, and executes SQL under read-only mode, row limiting, and query timeouts. |
| I need PostgreSQL performance tuning and health checks | Use Postgres MCP (crystaldba/postgres-mcp): it is a Community favourite with 3,402 GitHub stars and 89 stars gained per 30 days, its tool descriptions are graded A, and it provides tools for index recommendations, EXPLAIN plans, top queries, database health, and safe SQL execution. |
| I monitor production PostgreSQL with read-only analysis | Use MCP PostgreSQL Operations (call518/MCP-PostgreSQL-Ops): it exposes 34 tools for locks, WAL, replication, bloat, vacuum, query, and schema insights, its tool descriptions are graded A, and it had a commit 0 days ago. |
| I write PostgreSQL code and need documentation search | Use pg-aiguide (timescale/pg-aiguide): it enables semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation and provides installable Postgres Agent Skills. |
| I want simple read-only queries and schema inspection | Use MCP PostgreSQL Server (antonorlov/mcp-postgres-server): it exposes query, list_schemas, list_tables, and describe_table tools with writes disabled by default, and its npm package had 1,134 downloads in a typical week. |
| I need broad production operations with many built-in tools | Use MCPg - Production-grade PostgreSQL MCP Server (devopam/MCPg): it exposes 191 tools for catalog introspection, query execution, lock inspection, and index tuning, and it had a commit 0 days ago, though its tool descriptions are graded B. |
Top MCP servers for PostgreSQL
| Best for | Profile | ||||||
|---|---|---|---|---|---|---|---|
| 1 | AI coding assistants writing PostgreSQL code: it provides semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation plus installable Postgres Agent Skills. | Community favourite | 1,869 | +27 | 11 days ago | 80.4 | |
| 2 | PostgreSQL performance tuning and health checks: it exposes 9 tools for index recommendations, EXPLAIN plans, top queries, database health, and safe SQL execution. | Community favourite | 3,402 | +89 | 64 days ago | 79.3 | |
| 3 | Internal developer workflows against trusted PostgreSQL databases: it provides SQL query access through MCP-compatible clients and a natural language agent for forming SQL queries. | Steady | 236 | +10 | 20 days ago | 73.4 | |
| 4 | Local development and read-only production troubleshooting alongside coding agents: it executes SQL under read-only mode, row limiting, and query timeouts. | Community favourite | 3,639 | +125 | 5 days ago | 69.0 | |
| 5 | Monitoring PostgreSQL 12-18 in production with read-only analysis: it exposes 34 tools for locks, WAL, replication, bloat, vacuum, query, and schema insights. | Steady | 161 | +1 | today | 66.8 | |
| 6 | Prisma 8 projects on PostgreSQL: it installs per-workflow SKILL.md files that let Claude Code, Cursor and other agents drive schema and query changes. | Community favourite | 47,694 | +34 | 2 days ago | 66.3 | |
| 7 | Postgres schema, role, and index administration from an AI assistant: admin security mode unlocks schema, trigger, RLS, user, and index tools. | Abandoned but popular | 201 | +1 | 117 days ago | 64.7 | |
| 8 | Production Postgres operations alongside application queries: one server exposing 191 tools for catalog introspection, query execution, lock inspection, and index tuning. | Steady | 13 | +3 | today | 58.9 | |
| 9 | Already working in VS Code with configured database connections: its built-in MCP server exposes those connections, schemas, and query execution to agents. | Community favourite | 357 | +9 | 6 days ago | 57.7 | |
| 10 | Self-hosting a Postgres backend that needs RLS, auth, storage, serverless TypeScript functions, realtime, and RAG, with MCP access at /mcp or stdio. | Community favourite | 3,112 | -730 | 3 days ago | 57.7 | |
| 11 | Teams running several database engines who want one MCP client: it provides natural language queries and database management through the DBX manager. | Community favourite | 27,039 | +9,004 | today | 57.5 | |
| 12 | For connecting an MCP client to PostgreSQL: the README documents an npx command using --prebuilt=postgres --stdio, plus list_tables and execute_sql. | Community favourite | 16,673 | +259 | 3 days ago | 54.9 | |
| 13 | Read-only PostgreSQL analysis via DATABASE_URL: it exposes a query tool and table schema resources, with write operations disabled unless DANGEROUSLY_ALLOW_WRITE_OPS is enabled. | Abandoned but popular | 179 | -1 | 349 days ago | 51.2 | |
| 14 | Read-only natural-language querying of a PostgreSQL database from Claude Desktop: it runs read-only by default and takes connection details as CLI flags or HTTP headers. | Abandoned but popular | 940 | +6 | 215 days ago | 51.1 | |
| 15 | Teams that need AI agents to query PostgreSQL alongside other systems under governed permissions and schema discovery, using GraphJin's MCP server and natural-language built-in agent. | Community favourite | 3,213 | +52 | today | 51.0 | |
| 16 | Running several MCP clients against one shared memory: the README documents a streamable HTTP mode that lets multiple agents connect to a single server instance. | Steady | 115 | 0 | 21 days ago | 50.8 | |
| 17 | Existing Nhost projects: the CLI's built-in MCP server exposes six tools for schema, allowed GraphQL operations, Nhost management, and docs. | Community favourite | 9,357 | No snapshot history | today | 49.1 | |
| 18 | AI-assisted management of a hosted WhoDB platform project with PostgreSQL data: it exposes 155 tools for datasets, ontologies, apps, transforms, and AI providers. | Community favourite | 5,034 | +6 | today | 48.9 | |
| 19 | Team wikis that agents should read and write directly: it exposes 50 tools covering search, page CRUD, comments, attachments and Atlas doc generation. | Emerging | 82 | +19 | 3 days ago | 48.8 | |
| 20 | Read-only querying and schema inspection against an existing PostgreSQL database: it exposes query, list_schemas, list_tables, and describe_table tools with writes disabled by default. | Community favourite | 23 | 0 | 32 days ago | 48.6 |
The ranking, with the evidence
Each position is a weighted mean of adoption (40%), maintenance (24%), momentum (14%), tool description quality (13%) and trust (9%), multiplied by three attenuators: how directly the server is about PostgreSQL (named for it, declaring it, tagged with it, or merely mentioning it), whether its repository is still moving, and how much independent evidence of adoption it has. Open the score on any entry to see every number, including the ones marked ≈, which were imputed from the median of the other candidates rather than measured. The maintenance grade on each entry is mostly issue responsiveness, release recency and open security alerts rather than commits, so a recent commit beside a low grade is two different measurements rather than a contradiction.
- Abandoned but popular: People use it, but its default branch has stopped moving. Fine to keep running, risky to adopt.
- Community favourite: Widely adopted and still actively maintained.
- Dormant: Neither changing nor widely adopted. Here because it still matches the search.
- Emerging: Small audience, growing quickly, maintained. The bet with the most upside.
- Steady: Maintained, modest audience, no surprises in either direction.
Best for: AI coding assistants writing PostgreSQL code: it provides semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation plus installable Postgres Agent Skills.
The README documents Agent Skills installed with npx skills, a Claude Code plugin, and a public MCP endpoint at https://mcp.tigerdata.com/docs for semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation. Its MCP use depends on the hosted endpoint at https://mcp.tigerdata.com/docs.
GitHub stars1,869Stars / 30 days+27npm / typical week67PyPI / typical weekno attributed packageTools exposednever inspectedLast commit11 days agoCommits / 12 weeks25Maintenance gradeATool descriptionsNot gradedScore 80.4: show every number behind it
- Adoption85 / 100 · weight 40%
- GitHub stars82
- npm downloads19downloads show none of the weekday rhythm human traffic has; halved
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance97 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence85
- Momentum39 / 100 · weight 14%
- Stars gained, relative to size48
- Stars gained, absolute58
- npm download trend0
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 80.4
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 80.4
Best for: PostgreSQL performance tuning and health checks: it exposes 9 tools for index recommendations, EXPLAIN plans, top queries, database health, and safe SQL execution.
The server exposes 9 tools for schema and object inspection, query plan explanation, index recommendations, database health checks, top query reporting, and SQL execution, and its README documents standard input/output and server-sent events transports. Before selecting it, note that it requires database access credentials and either Docker or Python 3.12 or higher.
GitHub stars3,402Stars / 30 days+89npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposed9Last commit64 days agoCommits / 12 weeks1Maintenance gradeDTool descriptionsAScore 79.3: show every number behind it
- Adoption88 / 100 · weight 40%
- GitHub stars88
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance64 / 100 · weight 24%
- Last commit touching this server88
- Repository maintenance grade30
- Commit cadence40
- Momentum71 / 100 · weight 14%
- Stars gained, relative to size65
- Stars gained, absolute79
- npm download trendnot measuredno download history for the selected registry
- Tool quality76 / 100 · weight 13%
- Tool description quality68
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 79.3
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 79.3
Best for: Internal developer workflows against trusted PostgreSQL databases: it provides SQL query access through MCP-compatible clients and a natural language agent for forming SQL queries.
The README documents SQL query access to PostgreSQL through MCP-compatible clients and a natural language agent for forming SQL queries, plus client setup for CLI, web, Go chat, Claude Desktop, and Cursor. It is explicitly not for public-facing applications because it provides LLMs with read access to the entire database schema and data, so it should only be used for internal tools, developer workflows, or trusted environments.
GitHub stars236Stars / 30 days+10npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposednever inspectedLast commit20 days agoCommits / 12 weeks68Maintenance gradeATool descriptionsNot gradedScore 73.4: show every number behind it
- Adoption59 / 100 · weight 40%
- GitHub stars59
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum57 / 100 · weight 14%
- Stars gained, relative to size66
- Stars gained, absolute42
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 73.4
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 73.4
Best for: Local development and read-only production troubleshooting alongside coding agents: it executes SQL under read-only mode, row limiting, and query timeouts.
DBHub runs as a single MCP server connecting clients such as Claude Desktop, Claude Code, Cursor, and VS Code to PostgreSQL, MySQL, SQL Server, MariaDB, Oracle, and SQLite, exposing tools for executing SQL and searching schema objects, plus opt-in explain and health-check tools and custom tools defined in a TOML config. Worth knowing before choosing it: it is launched through npx or Docker with a connection string, its default tool surface is deliberately small, and the README points anyone needing approval flows, access control, data masking, or audit logging to Bytebase instead.
GitHub stars3,639Stars / 30 days+125npm / typical weekdownloads not countedPyPI / typical weekno attributed packageTools exposednever inspectedLast commit5 days agoCommits / 12 weeks94Maintenance gradeATool descriptionsNot gradedScore 69.0: show every number behind it
- Adoption89 / 100 · weight 40%
- GitHub stars89
- npm downloadsnot measurednpm names no repository for dbhub, so its downloads cannot be attributed
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum79 / 100 · weight 14%
- Stars gained, relative to size75
- Stars gained, absolute85
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 88.5
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 69.0
Best for: Monitoring PostgreSQL 12-18 in production with read-only analysis: it exposes 34 tools for locks, WAL, replication, bloat, vacuum, query, and schema insights.
MCP PostgreSQL Operations exposes 34 tools for PostgreSQL monitoring and analysis, covering server and database metadata, connections, locks, WAL and replication status, schemas, table relationships, table and index I/O, VACUUM and autovacuum activity, bloat, and configuration parameters. The main caveat is that advanced query analysis depends on the pg_stat_statements extension, with pg_stat_monitor optional, and the server documents read-only operations.
GitHub stars161Stars / 30 days+1npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposed34Last committodayCommits / 12 weeks20Maintenance gradeBTool descriptionsAScore 66.8: show every number behind it
- Adoption55 / 100 · weight 40%
- GitHub stars55
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance92 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade80
- Commit cadence85
- Momentum23 / 100 · weight 14%
- Stars gained, relative to size28
- Stars gained, absolute16
- npm download trendnot measuredno download history for the selected registry
- Tool quality81 / 100 · weight 13%
- Tool description quality73
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 66.8
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 66.8
Best for: Prisma 8 projects on PostgreSQL: it installs per-workflow SKILL.md files that let Claude Code, Cursor and other agents drive schema and query changes.
Its exposed tool list has not been measured, so what it does has to be read from its README: the installers scaffold a project, write prisma.config.ts, and drop a prisma-8.md primer plus one SKILL.md per workflow into .claude/skills, .cursor/skills, .agents/skills and .devin/skills so a coding agent can drive Prisma changes end to end. Before installing, note that it assumes Node.js 24 or newer and that Prisma 8 is a release candidate which may include breaking changes until 8.0.0 final, while Prisma 7 stays on the v7 branch.
GitHub stars47,694Stars / 30 days+34npm / typical weekdownloads not countedPyPI / typical weekno attributed packageTools exposednever inspectedLast commit2 days agoCommits / 12 weeks460Maintenance gradeBTool descriptionsNot gradedScore 66.3: show every number behind it
- Adoption100 / 100 · weight 40%
- GitHub stars100
- npm downloadsnot measurednpm names no repository for dev-env, so its downloads cannot be attributed
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance95 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade80
- Commit cadence100
- Momentum32 / 100 · weight 14%
- Stars gained, relative to size11
- Stars gained, absolute62
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 85.0
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 66.3
Best for: Postgres schema, role, and index administration from an AI assistant: admin security mode unlocks schema, trigger, RLS, user, and index tools.
The server exposes 18 tools covering schema, functions, triggers, indexes, constraints, row-level security, users and permissions, EXPLAIN-based query analysis, SELECT and mutation execution, export/import and cross-database copying, object comments, and monitoring. It starts in readonly mode, so mutations, DDL, role administration, filesystem import/export, and arbitrary SQL each need an explicit security mode, and destructive operations such as drops additionally require --allow-destructive.
GitHub stars201Stars / 30 days+1npm / typical week1.7KPyPI / typical weekno attributed packageTools exposed18Last commit117 days agoCommits / 12 weeks0Maintenance gradeDTool descriptionsBScore 64.7: show every number behind it
- Adoption77 / 100 · weight 40%
- GitHub stars58
- npm downloads69
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance57 / 100 · weight 24%
- Last commit touching this server88
- Repository maintenance grade30
- Commit cadence5
- Momentum21 / 100 · weight 14%
- Stars gained, relative to size25
- Stars gained, absolute15
- npm download trend19
- Tool quality63 / 100 · weight 13%
- Tool description quality55
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 64.7
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 64.7
Best for: Production Postgres operations alongside application queries: one server exposing 191 tools for catalog introspection, query execution, lock inspection, and index tuning.
It exposes 191 tools over stdio and HTTP/SSE covering schema introspection (tables, indexes, constraints, roles, grants, policies), SQL execution guarded by an AST allowlist, and DBA views such as pg_locks, pg_stat_io, and WAL archive status. Write capability is gated: the server runs read-only by default, with DDL, shell, and LISTEN/NOTIFY disabled until explicitly opted in, so a fresh install yields an inspecting agent rather than a mutating one.
GitHub stars13Stars / 30 days+3npm / typical weekShips no npm packagePyPI / typical week198mcpgTools exposed191Last committodayCommits / 12 weeks91Maintenance gradeATool descriptionsBScore 58.9: show every number behind it
- Adoption53 / 100 · weight 40%
- GitHub stars29
- PyPI downloads49
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum47 / 100 · weight 14%
- Stars gained, relative to size60
- Stars gained, absolute26
- PyPI download trendnot measuredno download history for the selected registry
- Tool quality66 / 100 · weight 13%
- Tool description quality57
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 69.3
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: modest but real audience
- 0.85
- Composite score
- 58.9
Best for: Already working in VS Code with configured database connections: its built-in MCP server exposes those connections, schemas, and query execution to agents.
DBCode is a VS Code extension for managing PostgreSQL, MySQL, SQL Server, and other databases, and it offers the option to run an MCP server that gives agents access to those databases, their schemas, and query execution. Choosing it assumes the extension is installed with a connection configured, since the MCP server runs from that extension rather than as a standalone package.
GitHub stars357Stars / 30 days+9npm / typical weekdownloads not countedPyPI / typical weekno attributed packageTools exposednever inspectedLast commit6 days agoCommits / 12 weeks69Maintenance gradeATool descriptionsNot gradedScore 57.7: show every number behind it
- Adoption64 / 100 · weight 40%
- GitHub stars64
- npm downloadsnot measurednpm names no repository for dbcode, so its downloads cannot be attributed
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum48 / 100 · weight 14%
- Stars gained, relative to size53
- Stars gained, absolute40
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 74.0
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 57.7
Best for: Self-hosting a Postgres backend that needs RLS, auth, storage, serverless TypeScript functions, realtime, and RAG, with MCP access at /mcp or stdio.
Butterbase documents an MCP server at /mcp on a self-hosted instance and a stdio interface via npx @butterbase/mcp, for managing a Postgres-backed backend with RLS, auth, storage, serverless TypeScript functions, durable objects, realtime, KV, AI gateway, and RAG. The HTTP MCP endpoint assumes a self-hosted instance, and the npm package for the stdio route is listed as not published.
GitHub stars3,112Stars / 30 days-730npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposednever inspectedLast commit3 days agoCommits / 12 weeks134Maintenance gradeCTool descriptionsNot gradedScore 57.7: show every number behind it
- Adoption87 / 100 · weight 40%
- GitHub stars87
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance89 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade55
- Commit cadence100
- Momentum0 / 100 · weight 14%
- Stars gained, relative to size0
- Stars gained, absolute0
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 74.0
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 57.7
Best for: Teams running several database engines who want one MCP client: it provides natural language queries and database management through the DBX manager.
The repository documents DBX as a desktop application, Docker image, CLI, built-in AI assistant, and MCP server, with the MCP component providing natural language queries and database management. What is not documented is its tool list: no inspection has been recorded, so the exact MCP tools it exposes remain unknown.
GitHub stars27,039Stars / 30 days+9,004npm / typical weekdownloads not countedPyPI / typical weekno attributed packageTools exposednever inspectedLast committodayCommits / 12 weeks4,596Maintenance gradeATool descriptionsNot gradedScore 57.5: show every number behind it
- Adoption100 / 100 · weight 40%
- GitHub stars100
- npm downloadsnot measurednpm publishes dbx from samt/node-dbx
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum100 / 100 · weight 14%
- Stars gained, relative to size100
- Stars gained, absolute100
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 95.8
- × relevance: the keyword is tagged here
- 0.60
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 57.5
Best for: For connecting an MCP client to PostgreSQL: the README documents an npx command using --prebuilt=postgres --stdio, plus list_tables and execute_sql.
MCP Toolbox for Databases is an open source MCP server whose README documents connecting MCP-compatible clients to PostgreSQL through an npx command using @toolbox-sdk/server, --prebuilt=postgres, and --stdio, alongside prebuilt generic tools such as list_tables and execute_sql. Before choosing it, note that the prebuilt route requires setting environment variables for the database connection, and the README points to the Prebuilt Tools Reference for those settings.
GitHub stars16,673Stars / 30 days+259npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposednever inspectedLast commit3 days agoCommits / 12 weeks221Maintenance gradeATool descriptionsNot gradedScore 54.9: show every number behind it
- Adoption100 / 100 · weight 40%
- GitHub stars100
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum69 / 100 · weight 14%
- Stars gained, relative to size51
- Stars gained, absolute97
- npm download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 91.5
- × relevance: the keyword is tagged here
- 0.60
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 54.9
Best for: Read-only PostgreSQL analysis via DATABASE_URL: it exposes a query tool and table schema resources, with write operations disabled unless DANGEROUSLY_ALLOW_WRITE_OPS is enabled.
Postgres MCP Server connects to PostgreSQL using a required DATABASE_URL, exposes a
querytool for executing SQL, and documentspostgres://tablesandpostgres://table/{schema}/{table}resources for table listing and schema/sample data retrieval; it also supports stdio and HTTP transports, with HTTP serving the MCP Streamable HTTP endpoint at/mcp. Before choosing it, note that write operations are disabled by default and require enabling DANGEROUSLY_ALLOW_WRITE_OPS.GitHub stars179Stars / 30 days-1npm / typical week207PyPI / typical weekno attributed packageTools exposednever inspectedLast commit349 days agoCommits / 12 weeks0Maintenance gradeDTool descriptionsNot gradedScore 51.2: show every number behind it
- Adoption64 / 100 · weight 40%
- GitHub stars56
- npm downloads49
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance32 / 100 · weight 24%
- Last commit touching this server42dated from the last commit on the default branch, re-read from GitHub at publication; github.com shows a push 60 days ago, which counts every ref; the stored date would have published 60 days ago
- Repository maintenance grade30
- Commit cadence5
- Momentum2 / 100 · weight 14%
- Stars gained, relative to size0
- Stars gained, absolute0
- npm download trend9
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 51.2
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 51.2
Best for: Read-only natural-language querying of a PostgreSQL database from Claude Desktop: it runs read-only by default and takes connection details as CLI flags or HTTP headers.
Universal DB MCP runs as an MCP server over stdio, SSE, or Streamable HTTP (plus a REST API), sending queries to PostgreSQL, MySQL, Redis and other databases, and defaults to read-only access. Before installing, note that it requires Node.js 20 or later and passes connection details as command-line flags or HTTP headers, and that PostgreSQL is one of many supported backends rather than a dedicated target.
GitHub stars940Stars / 30 days+6npm / typical week91PyPI / typical weekno attributed packageTools exposednever inspectedLast commit215 days agoCommits / 12 weeks0Maintenance gradeDTool descriptionsNot gradedScore 51.1: show every number behind it
- Adoption81 / 100 · weight 40%
- GitHub stars74
- npm downloads42
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance46 / 100 · weight 24%
- Last commit touching this server68
- Repository maintenance grade30
- Commit cadence5
- Momentum32 / 100 · weight 14%
- Stars gained, relative to size30
- Stars gained, absolute33
- npm download trend34
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 65.5
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 51.1
Best for: Teams that need AI agents to query PostgreSQL alongside other systems under governed permissions and schema discovery, using GraphJin's MCP server and natural-language built-in agent.
GraphJin exposes access to databases, files, source code, workflows, metadata, and security policy through GraphQL and MCP, and its README documents MCP tools including query_catalog, graphql_help, ask_graphjin_agent, gj_security, gj_runtime, and gj_watch. Before choosing it, note that the built-in agent depends on a model API key in ./.env, and guarded access includes source-mode access, query allow-lists, read-only boundaries, and policy-aware MCP tools.
GitHub stars3,213Stars / 30 days+52npm / typical week420PyPI / typical weekno attributed packageTools exposednever inspectedLast committodayCommits / 12 weeks522Maintenance gradeATool descriptionsNot gradedScore 51.0: show every number behind it
- Adoption92 / 100 · weight 40%
- GitHub stars88
- npm downloads28downloads show none of the weekday rhythm human traffic has; halved
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum46 / 100 · weight 14%
- Stars gained, relative to size51
- Stars gained, absolute70
- npm download trend8
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 85.0
- × relevance: the keyword is tagged here
- 0.60
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 51.0
Best for: Running several MCP clients against one shared memory: the README documents a streamable HTTP mode that lets multiple agents connect to a single server instance.
Ogham MCP is a Python server, distributed as the ogham-mcp package, that keeps agent memory in PostgreSQL or Supabase and retrieves it with hybrid search plus a typed-edge graph, speaking stdio or streamable HTTP to any MCP client. Before installing, note that it needs a database (Supabase, Neon, or self-hosted Postgres) and a configured embedding provider such as OpenAI, Ollama, Mistral, or Voyage, plus the schema migration that ogham init runs.
GitHub stars115Stars / 30 days0npm / typical weekShips no npm packagePyPI / typical week119ogham-mcpTools exposednever inspectedLast commit21 days agoCommits / 12 weeks36Maintenance gradeATool descriptionsNot gradedScore 50.8: show every number behind it
- Adoption58 / 100 · weight 40%
- GitHub stars52
- PyPI downloads44
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum0 / 100 · weight 14%
- Stars gained, relative to size0
- Stars gained, absolute0
- PyPI download trendnot measuredno download history for the selected registry
- Tool quality≈68 / 100 · weight 13%
- Tool description quality≈68tool descriptions not yet scored
- Built and inspected by Glamanot measurednever built and inspected by Glama
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 65.1
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 50.8
Best for: Existing Nhost projects: the CLI's built-in MCP server exposes six tools for schema, allowed GraphQL operations, Nhost management, and docs.
Nhost ships an MCP server inside its CLI that exposes six tools: get-schema, graphql-query, list, manage-graphql, read_page, and search. Using it requires the Nhost CLI and an Nhost project, since those tools read GraphQL schemas, run allowed GraphQL operations, manage local metadata and migrations, update Nhost Cloud project config, and search Nhost docs.
GitHub stars9,357Stars / 30 daysno snapshot historynpm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposed6Last committodayCommits / 12 weeks271Maintenance gradeATool descriptionsAScore 49.1: show every number behind it
- Adoption99 / 100 · weight 40%
- GitHub stars99
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum≈0 / 100 · weight 14%
- Stars gained, relative to sizenot measuredno snapshot history yet
- Stars gained, absolutenot measuredno snapshot history yet
- npm download trendnot measuredno download history for the selected registry
- Tool quality71 / 100 · weight 13%
- Tool description quality63
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 81.8
- × relevance: the keyword is tagged here
- 0.60
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 49.1
Best for: AI-assisted management of a hosted WhoDB platform project with PostgreSQL data: it exposes 155 tools for datasets, ontologies, apps, transforms, and AI providers.
WhoDB CLI exposes 155 tools, including whodb_mcp_setup guidance and whodb_platform_* actions for hosted platform resources such as datasets, ontologies, apps, transforms, and AI providers. No npm package is published and no PyPI package is attributed, so installation goes through the WhoDB CLI rather than a package manager.
GitHub stars5,034Stars / 30 days+6npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposed155Last committodayCommits / 12 weeks602Maintenance gradeATool descriptionsBScore 48.9: show every number behind it
- Adoption93 / 100 · weight 40%
- GitHub stars93
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum21 / 100 · weight 14%
- Stars gained, relative to size14
- Stars gained, absolute33
- npm download trendnot measuredno download history for the selected registry
- Tool quality66 / 100 · weight 13%
- Tool description quality57
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integrates100
- Weighted mean of the five
- 81.5
- × relevance: the keyword is tagged here
- 0.60
- × continuity: actively changing
- 1.00
- × evidence: widely adopted
- 1.00
- Composite score
- 48.9
Best for: Team wikis that agents should read and write directly: it exposes 50 tools covering search, page CRUD, comments, attachments and Atlas doc generation.
tela is a self-hostable, markdown-native wiki whose own backend serves MCP over /api/mcp, exposing 50 tools that search pages by ranked full-text and semantic retrieval, read and write pages, comments, attachments, spaces and share links, and trigger Atlas doc-generation runs over git repositories and Jira. Choosing it means running the stack yourself with Docker Compose and make, and knowing that pgvector-backed semantic search only works once an embedder is configured and the vector extension is enabled on the database.
GitHub stars82Stars / 30 days+19npm / typical weekShips no npm packagePyPI / typical weekno attributed packageTools exposed50Last commit3 days agoCommits / 12 weeks184Maintenance gradeATool descriptionsAScore 48.8: show every number behind it
- Adoption48 / 100 · weight 40%
- GitHub stars48
- npm downloadsnot measuredno npm package
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance100 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade100
- Commit cadence100
- Momentum81 / 100 · weight 14%
- Stars gained, relative to size100
- Stars gained, absolute52
- npm download trendnot measuredno download history for the selected registry
- Tool quality78 / 100 · weight 13%
- Tool description quality70
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 73.6
- × relevance: the keyword is declared here
- 0.78
- × continuity: actively changing
- 1.00
- × evidence: modest but real audience
- 0.85
- Composite score
- 48.8
Best for: Read-only querying and schema inspection against an existing PostgreSQL database: it exposes query, list_schemas, list_tables, and describe_table tools with writes disabled by default.
MCP PostgreSQL Server connects to local, Docker, RDS, Neon, Supabase, or SSH bastion PostgreSQL databases and exposes query, list_schemas, list_tables, describe_table, and execute tools over stdio. Before choosing it, note that it requires Node.js 20 or newer, and the execute tool refuses writes unless PG_ALLOW_WRITE is set to true.
GitHub stars23Stars / 30 days0npm / typical week1.1KPyPI / typical weekno attributed packageTools exposed5Last commit32 days agoCommits / 12 weeks16Maintenance gradeBTool descriptionsAScore 48.6: show every number behind it
- Adoption65 / 100 · weight 40%
- GitHub stars35
- npm downloads60downloads capped: not corroborated by repository audience
- Used through Glamanot measurednot used through Glama in the last 30 days
- Maintenance92 / 100 · weight 24%
- Last commit touching this server100
- Repository maintenance grade80
- Commit cadence85
- Momentum0 / 100 · weight 14%
- Stars gained, relative to size0
- Stars gained, absolute0
- npm download trend0
- Tool quality96 / 100 · weight 13%
- Tool description quality88
- Built and inspected by Glama100
- Trust100 / 100 · weight 9%
- License100
- Published by the vendor it integratesnot measurednot published by the vendor it integrates
- Weighted mean of the five
- 69.4
- × relevance: the keyword is dedicated here
- 1.00
- × continuity: actively changing
- 1.00
- × evidence: small audience
- 0.70
- Composite score
- 48.6
Questions people ask
How do I choose between DBHub and Postgres MCP (crystaldba/postgres-mcp)?
DBHub (bytebase/dbhub) is the more active general option: it had a commit 5 days ago, 94 commits on its default branch in the last 12 weeks, 3,639 GitHub stars, 125 stars gained per 30 days, and read-only mode, row limiting, and query timeouts. Postgres MCP (crystaldba/postgres-mcp) is PostgreSQL-specific and its tool descriptions are graded A, with 3,402 GitHub stars and 89 stars gained per 30 days, but its last commit was 64 days ago and it had 1 commit on its default branch in the last 12 weeks. Choose DBHub for a universal client with recent activity, and choose Postgres MCP (crystaldba/postgres-mcp) for performance tuning and health checks.
What does MCP PostgreSQL Operations cover for production monitoring?
MCP PostgreSQL Operations (call518/MCP-PostgreSQL-Ops) exposes 34 tools for locks, WAL, replication, bloat, vacuum, query, and schema insights, and it is intended for production PostgreSQL monitoring with read-only analysis. Its tool descriptions are graded A, it had a commit 0 days ago, and it has 20 commits on its default branch in the last 12 weeks. It is a Steady project with 161 GitHub stars and 1 star gained per 30 days.
What should I check before using MCPg for production operations?
MCPg - Production-grade PostgreSQL MCP Server (devopam/MCPg) exposes 191 tools for catalog introspection, query execution, lock inspection, and index tuning, so the surface area is large. Its tool descriptions are graded B, it had a commit 0 days ago and 91 commits on its default branch in the last 12 weeks, and it is a Steady project with 13 GitHub stars and 3 stars gained per 30 days. Inspect the tools before enabling broad production use.
Can I use pg-aiguide to query a PostgreSQL database?
pg-aiguide (timescale/pg-aiguide) is described as enabling semantic search across PostgreSQL, TimescaleDB, and PostGIS documentation to help AI coding assistants write better PostgreSQL code, plus installable Postgres Agent Skills. It is a Community favourite with 1,869 GitHub stars, 27 stars gained per 30 days, an npm package @tigerdata/pg-aiguide, 67 npm downloads in a typical week, and a commit 11 days ago. If you need SQL query execution against a database, look at DBHub or Postgres MCP (crystaldba/postgres-mcp) instead.
What is the risk with an Abandoned but popular PostgreSQL MCP server?
PostgreSQL MCP Server (HenkDz/postgresql-mcp-server) is labeled Abandoned but popular, had its last commit 117 days ago, and had 0 commits on its default branch in the last 12 weeks, even though its npm package had 1,684 downloads in a typical week and it exposes 18 tools. If you need current maintenance, prefer a server with recent commits such as DBHub, MCP PostgreSQL Operations, or MCPg. The description says its admin security mode unlocks schema, trigger, RLS, user, and index tools, which matters if you do not need those administrative capabilities.