Browse tables, collections and keys, and run SQL, mongosh or Redis commands inline — for PostgreSQL, MySQL/MariaDB, MongoDB and Redis. No Adminer, no pgweb, no exposed port. Built into the binary, gated by RBAC, every query audited.
Every database service gets a Data tab. PostgreSQL (incl. PostGIS/TimescaleDB) and MySQL/MariaDB use native sqlx drivers; Redis/Valkey use the native redis client; MongoDB is driven through mongosh inside the container, so no heavy native Mongo driver bloats the binary.
Walk the schema/table tree, inspect columns, types, primary keys and indexes, and page through rows with a total count. For MongoDB, browse collections and paginated documents as EJSON; for Redis, scan keys with type-aware value views and TTL. The query runner sits below, collapsed by default — browse first.
Coolify and Dokploy make you run a separate Adminer/pgweb container or connect from outside. Pier connects server-side over the pier-net Docker network, so a private database never has to be published to the internet to inspect it.
Reads need Viewer, writes need Editor — give a teammate read-only access to production data without write rights. Every runner execution lands in the db_query_log table: who ran what, against which database, status, row count and duration.
Deploy any database from the catalog (PostgreSQL, MySQL/MariaDB, MongoDB, Redis/Valkey) and open its resource page.
Pier detects the engine and shows the matching browser — a schema/table tree for SQL, a collection browser for MongoDB, or a key browser for Redis.
Select a table to see columns, types, primary keys and indexes, then page through its rows. For MongoDB, page through documents; for Redis, open a key to see its value and TTL.
Open the SQL Runner (or Mongo Shell / Redis command box), run a statement, and read the result inline. Reads return a grid capped at 1,000 rows; writes report affected rows. Requires the Editor role.
PostgreSQL (including PostGIS and TimescaleDB), MySQL and MariaDB, MongoDB, and Redis/Valkey. SQL engines and Redis use native Rust drivers; MongoDB is driven through the mongosh already in the official image.
It's an 80% tool: browse the schema, eyeball rows, run a quick SELECT, fix a record, inspect a key — without leaving the dashboard. It deliberately doesn't try to replace a full client for ER diagrams, query-plan profiling or schema migrations.
Browsing is read-only and needs only Viewer. The runners require Editor and are fully audited in db_query_log. SQL reads are capped at 1,000 rows with a 15-second statement timeout, and identifiers in the browser are validated against information_schema and engine-quoted.
No. Pier core runs on the host and reaches the container over the pier-net Docker network (with a 127.0.0.1 fallback for published ports), so a private database stays private while you inspect it.
One command installs Pier on any Ubuntu or Debian VPS.
curl -fsSL https://pier.sh/install | sudo bash