Dynobase

Dynobase reads credentials from ~/.aws, the same place the AWS CLI and SDKs look. Profiles with MFA, SSO based identities, and aws-vault setups show up in a sidebar, and you can jump between accounts and regions without closing a window. Short lived and long lived credentials both work.

You name a table and the builder handles the rest: Query versus Scan, which index, how the key condition should read. PartiQL and SQL statements are accepted too, which helps when you already know the syntax. Every query exports as a CLI command or an SDK snippet.

Items open as editable JSON, maps and lists included. You change a nested attribute in place and save, and this DynamoDB editor writes only what changed. For anyone who has hand built an UpdateExpression, that removes most of the syntax errors.

Update, delete, and transaction operations get a form based builder that outputs production ready code. It matters most for conditional writes and multi item transactions, where a typo in a condition expression turns into a silent no-op.

Dynobase AI turns a plain description into a DynamoDB snippet for JavaScript, TypeScript, Python, Go, or Rust. You describe the operation, review the output, and paste it into your project. Treat it as a first draft, since the generated code still needs a read before it ships.

If your table holds embeddings, you can query the vector index with text, a pasted embedding, or an existing item, then inspect ranked matches. That saves writing a throwaway script every time you want to check retrieval quality.

The explorer maps nested paths, missing attributes, and mixed types across loaded results or a bounded sample. JSON path filters then pull a nested value into the data grid as a saved column, so a table with deep documents stays readable.