Features

Everything in the toolbox.

The full inventory: the headline features, the supporting cast, and the small stuff that makes the hours pleasant.

02

Safe in both directions

Prod is red here. The tool assumes you are one keystroke from an expensive mistake, and guards both directions.

Environment safety

A four-phase engine classifies prod, staging, and dev, with color-coded badges, custom rules, and overrides.

Explore →

Pending changes

Every edit is staged. Review the full diff, revert any change, and commit only when it is right, with prod labeled in red the whole way.

Explore →

No read on prod until you ask

Opening a table on a prod-classified connection issues no automatic read at all. Everywhere else, auto-scan loads a single bounded page, never the whole table.

Cost controls

Read limits you set, and a warning before anything bills.

Three-tier settings
Page size, batch size, and auto-scan set globally, per workspace, and per table.
Cancel anything mid-flight
Every expensive operation is cancellable while it runs.
Warned before it bills
S3 scan exports and stream-history drains name the billable throughput and throttling risk first.
Wasteful scans flagged
Filters discarding most of what was read get flagged in command history.

03

Your data model, understood

Single-table design, recovered from the data already in your table. No ORM required.

Bring your own single table design

Inference reads the keys and attributes already in your table and recovers the entities, the item collections they form, and the access patterns those collections support. No schema file, no library, no annotations.

Explore →

Evidence, not magic

Ranked against 13 known single-table design conventions, from The DynamoDB Book to PynamoDB, modyne, and others. You see what matched, what is missing, and where the table conflicts with itself. Inference is deterministic and shows its work.

Schema-aware query

Import an ElectroDB or DynamoDB Toolbox schema to skip the inference step. Either way you land in the same query surface: pick an entity and a pattern, and the real key condition is built underneath.

Explore →

AI, scoped and optional

It only ever suggests names, never structure, and it is off by default. AI is opt-in and sends your schema, never your data: table structure only, through your own provider, or fully on-device on Apple Silicon.

Simple one-entity tables do not get access patterns and do not need them. The feature shows up where the pain is.

05

Yours stays yours

Native code, local credentials, every platform, two languages.

Six credential types

IAM keys, MFA, IAM Identity Center SSO, AWS Login with passkeys, credential process, and local endpoints, with assume-role chains on top of any of them, across as many profiles and regions as you have.

Explore →

Credentials stay on your machine

Reads your existing ~/.aws profiles in place and stores any credentials it manages in your OS keyring. Nothing leaves your machine.

Cross-platform native

Built with Rust and Tauri. macOS Universal (notarized), Windows with Azure Trusted Signing, Linux AppImage, Snap, and deb. Available in English and 日本語.

And the small stuff

Export tables to S3Command palette ⌘PMulti-tab workspaceTab groupsCustomizable hotkeysCapacity tracking per queryCost controls per tableOn-device AI on Apple SiliconSample mode, no AWS accountTable prefix filteringCode snippets libraryIn-app feedback

Try all of it, free for 14 days.

Try Free for 14 Days