Show HN: DynoTable – The DynamoDB GUI with real SQL and your AI agent
DynoTable is a local-first desktop client for Amazon DynamoDB that enables real SQL queries, visual joins, and an AI agent powered by AWS Bedrock—all using the user’s own credentials without data leaving their machine. It introduces Smart Tables, allowing users to visually define relationships between DynamoDB tables and perform live, read-only joins without schema migration or ETL. The SQL Workbench supports full SELECT statements with GROUP BY, COUNT, SUM, AVG, HAVING, and multi-table joins, a
Analysis
TL;DR
- DynoTable is a local-first desktop client for Amazon DynamoDB that enables real SQL queries, visual joins, and an AI agent powered by AWS Bedrock—all using the user’s own credentials without data leaving their machine.
- It introduces Smart Tables, allowing users to visually define relationships between DynamoDB tables and perform live, read-only joins without schema migration or ETL.
- The SQL Workbench supports full SELECT statements with GROUP BY, COUNT, SUM, AVG, HAVING, and multi-table joins, all executed locally on fetched data to minimize DynamoDB cost and latency.
- An integrated AI agent reads schemas, drafts writes as reviewable diffs, and operates entirely within the user’s AWS account via Bedrock, ensuring no third-party access to data.
- External agents like Claude Code, Cursor, or Codex can connect via MCP (Model Context Protocol) with scoped permissions and staged commits, enhancing developer workflow while maintaining security.
Why It Matters
This tool addresses critical pain points in working with DynamoDB: its lack of native SQL support, difficulty in querying across multiple tables, and limited AI integration without compromising data sovereignty. By bringing SQL-like capabilities and secure, local AI assistance directly to developers’ machines, DynoTable lowers the barrier for complex analytics and automation on NoSQL workloads while aligning with enterprise security and compliance needs.
Technical Details
- Local-First Architecture: All data processing—including SQL execution, joins, and AI inference—occurs client-side; only raw requests are sent to AWS services (DynamoDB, Bedrock) using the user’s IAM roles or SSO credentials.
- Smart Tables Visual Join Canvas: Users draw edges between tables based on PKs or GSIs; DynoTable materializes joined views as projections that recompute instantly upon edits, without altering source data.
- SQL Workbench with PartiQL Compatibility: Supports standard SQL aggregations and joins over DynamoDB data by fetching minimal required rows via optimal access patterns (Query > Scan), then computing results locally.
- AI Agent Integration via Bedrock: Leverages Amazon Bedrock models through the user’s AWS account; prompts, schema, and table content never leave the user’s environment, with write operations requiring explicit approval via diff review.
- MCP Agent Connectivity: Exposes a local MCP server enabling external LLM agents (e.g., Cursor, Claude Code) to interact with DynamoDB under user-defined scopes (read-only, staging, full), with all changes held in a staging area until committed manually.
Industry Insight
The rise of local-first, credential-preserving AI tools for cloud databases signals a shift toward developer-controlled, privacy-aware infrastructure—especially important as organizations adopt more generative AI workflows. DynoTable’s model of combining SQL abstraction over NoSQL stores with secure, extensible agent interfaces sets a precedent for how database clients should evolve: not just as viewers, but as intelligent, collaborative platforms that respect both performance constraints and data governance policies. This approach could become standard for other managed services where direct API access lacks analytical depth or safe AI augmentation.
Disclaimer: The above content is generated by AI and is for reference only.