AI Agent Hub
Back to plugins
🖥️

dsh-chatbi

Client Updated 2026.08.24

Run the following command in DeepSeek Harness:

dsh plugin install Huauauaa/dsh-chatbi

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install Huauauaa/dsh-chatbi in DeepSeek Harness to install this plugin; the source is at https://github.com/Huauauaa/dsh-chatbi .

About this plugin

Standing up a real ChatBI stack usually means spinning up a Java backend, wiring database drivers, and stitching MCP endpoints together — a chain that is long to build and costly to maintain. dsh-chatbi collapses all of that into a single local artifact: every connection, tool definition, and execution record lives in one local JSON store. SQLite runs through Node's built-in node:sqlite module, while MCP and API tools are dispatched over plain HTTP to whatever endpoints you configure. No external JVM, no sidecar container.

On the agent side the plugin exposes a tight set of structured tools — chatbi_sql_execute for local SQL, chatbi_mcp_call for registered MCP capabilities, chatbi_api_call for API requests, plus chatbi_status and the chatbi_list_* family for introspection. Paired with the Settings panel's five tabs (Overview, Databases, MCP Tools, API Tools, SQL Tools), you get a small visual console for managing connections and tool templates, letting the DSH agent fetch data and invoke external tools conversationally instead of forcing you to hand-write HTTP calls or SQL scripts every time.

dsh-chatbi is aimed at developers already running DSH agents who want a lightweight BI and tool-orchestration layer without maintaining a separate backend service. If your workflow touches local SQLite queries, calls across multiple MCP servers, or chains a fixed set of APIs, this plugin folds those scattered pieces into one cohesive agent toolchain and keeps the whole stack local-first and fully self-contained.

Use Cases

  • Run local SQLite queries conversationally via the DSH agent and get structured results back
  • Manage multiple MCP server tools uniformly and invoke them on demand through the agent
  • Wrap a fixed set of API calls into agent-callable tools, eliminating hand-written HTTP scripts

Best For

  • Developers already running DSH agents who need a lightweight BI and tool-orchestration layer
  • Engineers orchestrating multiple external tools (MCP/API/SQL) without maintaining a separate backend
  • Solo developers who prefer a local-first, fully self-contained toolchain