pgview

pgview is a lightweight, keyboard-driven PostgreSQL browser for the terminal. Built in Go using the same TUI framework as k9s, it lets you connect to any PostgreSQL-compatible endpoint — direct host, pgBouncer, RDS Proxy, or SSH tunnel — and navigate tables, inspect schemas, filter rows, edit data, and run arbitrary SQL queries, all without leaving your terminal.

Features

Installation

Prerequisites

Go 1.22 or later — go.dev/dl

Download a binary

Pre-built binaries are available on the Releases page for linux and macOS (amd64 and arm64).

# macOS arm64 example
curl -L https://github.com/sibasismukherjee/pgview/releases/latest/download/pgview-darwin-arm64 \
  -o /usr/local/bin/pgview
chmod +x /usr/local/bin/pgview

Build from source

git clone https://github.com/sibasismukherjee/pgview.git
cd pgview
make build        # produces ./pgview

Install to PATH

make install      # go install — places binary in $(go env GOPATH)/bin

Connecting

# Interactive prompts for all connection details
pgview

# Host and port with explicit credentials
pgview -url myproxy.internal:5432 -username myuser -dbname mydb

# Full DSN (password embedded in URL)
pgview -url "postgres://myuser:secret@localhost:5432/mydb?sslmode=disable"

# Disable SSL for local development
pgview -url localhost -username postgres -dbname mydb -sslmode disable

CLI Flags

Table List

The default view after connecting. Lists all tables and views in the database, grouped by schema.

 pgview              │  <↵> view  <d> schema  <i> stats              │  Tables
 admin@mydb · local  │  </> search  <r> refresh  <e> SQL  <q> quit  │  public
──────────────────────────────────────────────────────────────────────────────────────
  schema    table                   type
  public    orders                  BASE TABLE
  public    products                BASE TABLE
▶ public    customers               BASE TABLE
  reporting monthly_totals          VIEW

Fuzzy Table Search

Press / from the table list to open a full-screen fuzzy finder that searches across all schemas simultaneously. Results update as you type.

 pgview              │  <↵> open  <Esc> cancel  │  <↑↓> navigate  │  type to filter all schemas
 admin@mydb · local  │
──────────────────────────────────────────────────────────────────────────────────────
 / ord
──────────────────────────────────────────────────────────────────────────────────────
   public.orders
   public.order_items
   reporting.daily_order_summary

Matched characters are highlighted in blue. Schema names shown in muted gray, table names in white. Results are ranked by match quality — consecutive character runs and word-boundary hits score higher than scattered matches.

Data View

Shows rows of the selected table, 200 per page. Type-aware cell colouring: numbers, booleans, UUIDs, timestamps, JSON, and NULL each have a distinct colour.

 pgview              │  <Esc> back  <g> top  <G> bottom  │  <n>/<p> page  │  </> filter  │  Data  public.customers
 admin@mydb · local  │  <d> schema  <f> row view/edit  <E> export  <i> stats  │  42 rows  ~1.2K est · PK: id
──────────────────────────────────────────────────────────────────────────────────────────────────────────────
  id    name              status    created_at           tags
▶ 1     Alice Johnson     active    2024-01-15 09:23:11  {platform,growth}
  3     Carol White       active    2024-03-19 11:02:44  {platform,api}
  7     Eve Martinez      active    2024-05-01 16:14:09  {growth}

 WHERE "status"::text ILIKE 'active'

Filter DSL

Press / in the data view to open the filter prompt. Terms are whitespace-separated and AND-ed together.

# Show only active rows with a specific tag
/ status=active tags=platform

# Date range on a timestamp column
/ created_at>=2024-01-01 created_at<2025-01-01

# Free-text search across all columns
/ johnson

Export to CSV / JSON

Press E (Shift+E) from the data view to start an interactive two-step export flow. Works for both table browse and SQL query results.

NULL values are written as an empty string in CSV and as null in JSON.

Row Viewer & Editor

Press f in the data view to open a full-screen two-column table showing every column of the selected row alongside its value.

 Row Viewer  <e>/<↵> edit  <Ctrl+S> save  <Esc> close  · public.customers · row 1
─────────────────────┬──────────────────────────────────────────────────────────
 Column              │ Value
─────────────────────┼──────────────────────────────────────────────────────────
 id                  │  1
 name                │  Alice Johnson
▶ status             │  active                              (edited)
 created_at          │  2024-01-15 09:23:11
 tags                │  {platform,growth}
─────────────────────┴──────────────────────────────────────────────────────────

 1 unsaved change(s) — Ctrl+S to save, Esc to discard

 edit status  ▏ inactive▌

Modified fields are highlighted in teal with an (edited) marker. The footer shows a count of unsaved changes. Type NULL (any case) to set a field to SQL NULL.

Schema Browser

Press d from the table list or data view to open a four-tab panel for the selected table. Navigate tabs with 1–4 or Tab / Shift+Tab.

 pgview              │  <1> Columns  <2> Indexes  <3> Constraints  <4> DDL  │  Schema  public.customers
 admin@mydb · local  │  <Tab> next tab  │  <↵> view data  <e> SQL  <Esc> back  <q> quit
─────────────────────────────────────────────────────────────────────────────────────────
 [1] Columns   [2] Indexes   [3] Constraints   [4] DDL
──────────────────────────────────────────────────────────
  Column       Type                    Nullable   Default
  id           bigint                  NOT NULL
  name         character varying(120)  NOT NULL
▶ status       text                    NULL       'active'
  created_at   timestamp with tz       NOT NULL   now()
  tags         text[]                  NULL

The DDL tab (4) shows a reconstructed CREATE TABLE statement using accurate type names from pg_catalog.format_type, inline constraints, and standalone CREATE INDEX statements for non-primary indexes:

 [1] Columns   [2] Indexes   [3] Constraints   [4] DDL
──────────────────────────────────────────────────────────
 CREATE TABLE "public"."customers" (
   "id"          bigint  NOT NULL  DEFAULT nextval('customers_id_seq'),
   "name"        character varying(120)  NOT NULL,
   "status"      text  DEFAULT 'active',
   "created_at"  timestamp with time zone  NOT NULL  DEFAULT now(),
   "tags"        text[],
   CONSTRAINT "customers_pkey"  PRIMARY KEY  (id),
   CONSTRAINT "customers_status_check"  CHECK  (status = ANY (ARRAY['active','inactive']))
 );

 CREATE INDEX idx_customers_status ON public.customers USING btree (status);

SQL Editor

Press e from any view to open a full-screen SQL editor. When opened from the data view it is pre-filled with the last executed query.

 pgview              │  <Ctrl+E> run  <Tab> complete  <Ctrl+L> clear  <Esc> cancel  │  SQL Editor
 admin@mydb · local  │  <Ctrl+R> history  <Ctrl+T> templates                              │
──────────────────────────────────────────────────────────────────────────────────────────
  SELECT *
  FROM "public"."customers"
  WHERE status = 'active'
  LIMIT 100▌

Schema-aware Tab completion

Completion is clause-sensitive. Pressing Tab prioritises suggestions based on context:

1. Type-matched operator when the previous token is a known column in WHERE/ON/HAVING — LIKE/ILIKE for text, >=/<= for timestamps, IS TRUE/IS FALSE for booleans, = ANY( for arrays, ->/->> for JSON
2. Table name in FROM/JOIN context
3. Column name in SELECT/WHERE/ORDER BY context
4. SQL keyword or table name as fallback

Templates Panel

Ctrl+T opens a panel on the left side of the SQL editor. Templates are pre-filled with the current table's real column names, quoted identifiers, and primary-key column.

  Templates                  │
  ── Query ─────────────     │  SELECT *
   SELECT *                  │  FROM "public"."customers"
   SELECT cols               │  LIMIT 100▌
   SELECT WHERE              │
   COUNT                     │
  ── Write ─────────────     │
   INSERT                    │
   UPDATE                    │
   DELETE                    │
   UPSERT                    │
  ── DDL ───────────────     │
   ADD COLUMN                │
   DROP COLUMN               │
   CREATE INDEX              │
   ANALYZE                   │
   TRUNCATE                  │

Use ↑/↓ to navigate (category headers are skipped), Enter to load the template into the editor, Esc to return focus to the editor.

Query History

Ctrl+R in the SQL editor opens a navigable history panel showing the last 50 queries, most-recent-first. Press Enter to load a query back into the editor, or Esc to close without loading.

Mouse & Touchpad

All table views (table list, data view, schema browser) support scroll gestures:

The SQL editor's text area handles its own scroll natively and is unaffected by the global capture.

Full Keyboard Reference

Table List

Fuzzy Search

Data View

Row Viewer / Editor

Schema Browser

SQL Editor

Makefile Targets

Contributing

Contributions are welcome. The short version:

1. Search or open an Issue first
2. Fork → branch from main → PR
3. Tests and lint must pass — CI runs automatically on every PR

See CONTRIBUTING.md for the full guide, CODE_OF_CONDUCT.md for community standards, and SECURITY.md for reporting vulnerabilities privately.

License

pgview is released under the MIT License. Use it freely in personal and commercial projects.