Tako MCP

Tako MCP is a simple MCP server that queries Tako and returns real-time data and visualization

Available Tools

search_tako

Takes a query to search Tako and the web to get real-time data and visualization. Returns embed, webpage, and image url of the visualization with relevant metadata such as source, methodology, and description.

upload_file_to_visualize

Takes a base64 encoded file as an input and uploads it to Tako to use for visualization

*If you call this tool with a big file, it may consume a large number of tokens and will be very slow. If you want to test visualizing bigger files though Tako, visit our playground

visualize_file

Use the file_id from upload_file_to_visualize and visualize the file. Returns embed, webpage, and image url of the visualization

visualize_dataset

Takes a Tako Data Format data and visualize. Returns embed, webpage, and image url of the visualization

Available Prompts

generate_search_tako_prompt

Prompt to assist the client to format query and search Tako using search_tako tool

generate_visualization_prompt

Prompt to assist the client to transform the data into Tako Data Format and visualize using visualize_dataset tool

Quickstart

Get your API key

Access Tako Dashboard and get your API key.

Installing via Smithery

To install tako-mcp for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @TakoData/tako-mcp --client claude

Add Tako MCP to Claude Desktop

Add the following to your .cursor/mcp.json or claude_desktop_config.json (MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json) ```json Python { "mcpServers": { "takoApi": { "command": "uv", "args": [ "--directory", "/path/to/tako/mcp", "run", "main.py" ], "env": { "TAKO_API_KEY": "" } } } }


## Example:
### 1. Use the prompt from Tako MCP Server `generate_search_tako_prompt`
The prompt will guide the model to generate optimized query to search Tako
### 2. Add your text input 
Add an input text to generate the prompt
> "Compare Magnificent 7 stock companies on relevant metrics."
### 3. Add a prompt to the chat 
Add additional instructions to the chat prompt
> Write me a research report on the magnificent 7 companies. Embed the result in an iframe whenever necessary
### 4. Checkout the result
  * [Claude Response](https://claude.ai/share/0c39e0c3-0811-486e-8f0b-92c8d5e05bc8)
  * [Generated Report](https://docs.trytako.com/documentation/integrations-and-examples/claude-generated-report)


## Environment Variables
### `ENVIRONMENT` 
Options:
- `remote` - If you're running a remote MCP server
- `local` - If you're running a local MCP server

### `TAKO_API_KEY`
- Your Tako API key, access it from [Tako Dashboard](https://trytako.com/dashboard)

## Testing Remote MCP
Start inspector and access the console

npx -y npx @modelcontextprotocol/inspector@latest


Start Tako MCP Server on remote mode

ENVIRONMENT=remote TAKO_API_KEY= uv run main.py

In inspector console, add the url `https://0.0.0.0:<port>/mcp/` and click connect

Select the `Tools` tab, and click `ListTools`. 

Select `search_tako` and test a query


## Deploying it on render
Since we use uv Render uses pip, we have to build a requirements.txt

uv pip compile pyproject.toml > requirements.txt ```