Local-first · Desktop-native

A database workbench
for humans & AI agents.

Wake is a fast, schema-aware desktop IDE for Postgres and the systems around it. Your queries, credentials, and data stay on your machine — every statement is reviewed before it runs.

macOS builds ship first. Windows and Linux in the queue. View all releases.

01 — Workbench

Query like you mean it.

A real editor with statement-level execution, Explain, history, and result inspection. No web tab to refresh, no session timeout, no surprise CSV truncation. Multi-engine notebooks let you join a Postgres query, a Redis lookup, and a chart in the same canvas.

  • Statement classification — SELECT vs DDL vs DML, color-coded confirmations
  • Per-connection history, scoped and searchable
  • Export to CSV, TSV, SQLite, INSERT scripts, Parquet
  • Notebook cells across SQL, Markdown, Redis, Mongo, Kafka tail, and Chart
Wake — Notebook ← Wake / Notebooks Untitled notebook saved ▶ Run all
SQL Photos (postgres) ▾ ▶ Run
1 select * from gear.cameras limit 10;
+ SQL · + Markdown · + Redis · + Mongo · + Kafka tail · + Chart
SQL Photos 2 (postgres) ▾ ▶ Run
1 select * from gear.cameras limit 2;
+ SQL · + Markdown · + Redis · + Mongo · + Kafka tail · + Chart
CHART Source: SQL cell #2 ▾
Pick a source SQL cell, X column, and Y column.
+ SQL · + Markdown · + Redis · + Mongo · + Kafka tail · + Chart
REDIS My Redis (redis) ▾ ▶ Run

02 — Schema-aware

Know the shape of your data.

Wake reads your schema, indexes, roles, settings, and replication state, and surfaces the parts that matter. Anti-pattern detection flags the things you'd otherwise find in production at 2am — oversized varchars, low-cardinality text columns, missing indexes on foreign keys.

  • Tables, columns, indexes, foreign keys, constraints, triggers
  • Replication lag, role grants, server settings, locks
  • Anti-pattern detection driven by pg_stats — no extra load
  • Per-engine status: Redis, Mongo, Kafka, RabbitMQ
gear.lenses owner: jhinds · 24 kB
COLUMNS INDEXES CONSTRAINTS TRIGGERS PERMISSIONS STATS COL STATS ANTI-PATTERNS 3 SECURITY ACTIVITY DDL
ColumnTypeNullableDefaultPK
lens_id integer NO nextval('gear.lenses_lens_id_seq')
manufacturer_id integer YES
model character varying(100) NO
focal_length_min integer YES
focal_length_max integer YES
max_aperture numeric(3,1) YES
mount character varying(30) YES
anti-pattern model · character varying(100) — avg width 14. Oversized declared length.

03 — Agents

Built for AI that actually touches your database.

LLMs are excellent at writing SQL and catastrophic at owning the consequences. Wake puts a typed, schema-aware layer between an agent and your production database — and a human review step in front of every write.

Assistant · proposed statement awaiting review
- delete from events where created_at < now() - interval '30 days';
+ delete from events
+ where created_at < now() - interval '30 days'
+   and tenant_id = $1
+ limit 10000;
targetproduction-pg / public.events
est. rows~1.2M
classificationDML · destructive
Approve & run Dry-run Reject

Schema as context, not guesswork.

Agents see your real tables, columns, types, and constraints. No hallucinated joins, no fabricated columns.

Reviewed before it runs.

Every statement is classified — SELECT, DDL, DML — and routed through an approval surface you can see and audit.

Local execution, local credentials.

Your database connection strings never leave the desktop. Agents propose; Wake executes locally against your network.

04 — Coverage

Postgres at the center. The systems around it, in the same workspace.

Today

A workbench engineers reach for daily.

Fast SQL, real schema inspection, exports, history. Local, native, quick.

Next

A control plane for data workflows.

Guardrails, observability, and policy around every human and agent that touches your data.

Get Wake

Install the latest build.

Signed builds are published through GitHub Releases. macOS first; Windows and Linux next.