tally
Tally Prime MCP (Model Context Protocol) Server implementation to feed Tally ERP data to popular LLM like Claude, ChatGPT supporting MCP
claude mcp add --transport stdio dhananjay1405-tally-mcp-server node D:\Software\Tally MCP Server\dist\index.mjs \ --env PORT="9000" \ --env TALLY_HOST="localhost" \ --env TALLY_PORT="9000" \ --env TALLY_XML_PORT="9000"
How to use
This MCP server implements the Tally Prime MCP (Model Context Protocol) interface and exposes a set of tools that enable an LLM (such as Claude or ChatGPT) to query and retrieve data from Tally Prime. The available tools support common accounting data extractions and summaries, including lists of masters, chart of accounts, trial balance, balance sheet, profit & loss, and ledger-level views. Each tool accepts specific inputs (like targetCompany, dates, and identifiers) and returns structured outputs that the LLM can incorporate into responses or use for follow-up questions. To use the server with an MCP client, point the client to the local MCP server (for example, http://localhost:9000 or the configured host/port) and invoke the desired tool with the required arguments. The server will communicate with Tally Prime over the configured XML port, fetch the data, and return the results in a machine-friendly format suitable for LLM reasoning and chain-of-thought prompts.
How to install
Prerequisites:
- Tally Prime (Silver/Gold) installed
- Node.js installed (recommended version as per Tally MCP requirements)
- XML Port of Tally Prime enabled (Settings > Connectivity > Client/Server configuration) with: TallyPrime acts as Server, Port = 9000
Install steps:
- Download the Tally MCP Server package zip from the official resource link provided in the project documentation
- Extract the zip to a local directory, e.g. D:\Software\Tally MCP Server
- Ensure the XML port is enabled in Tally Prime as described above
- Install dependencies (if any) by running: npm install
- Run the MCP server using Node.js: node dist/index.mjs
- Verify the server is listening on port 9000 and accessible from your MCP client (Claude Desktop, Perplexity, or ChatGPT integrations) by initiating a simple health check or a sample tool request
Note: If you deploy in a non-Windows environment or inside a container, adjust the command path accordingly (e.g., node dist/index.mjs or an equivalent entry point).
Additional notes
Tips and common issues:
- Ensure Tally Prime XML port is not blocked by firewall and is accessible from the machine running the MCP server.
- The educational version of Tally Prime may limit data ranges; use a full (non-educational) license to avoid incomplete responses.
- When using Claude Desktop or Perplexity Desktop, consider saving the MCP server configuration in the respective config files (e.g., claude_desktop_config.json) with the correct path to the server's entry point.
- If you see issues with path separators on Windows, remember to escape backslashes in JSON strings (e.g., D:\Software\Tally MCP Server\dist\index.mjs).
- If the MCP client reports missing ledgers or masters, you may need to run list-master first to populate available options, or adjust input values to closest matching names.
- The server communicates with Tally Prime via the XML Port (Port 9000). Ensure that port mapping remains consistent across local and cloud deployments.
Related MCP Servers
iterm
A Model Context Protocol server that executes commands in the current iTerm session - useful for REPL and CLI assistance
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
editor
MCP Server for Phaser Editor
DoorDash
MCP server from JordanDalton/DoorDash-MCP-Server
mcp
MCP сервер для автоматического создания и развертывания приложений в Timeweb Cloud