import * as fs from "fs/promises";

const schema = JSON.parse(await fs.readFile("./openapi.json", "utf8"));

// Remove local development server
schema.servers = schema.servers.filter((s) => !s.url.includes("//127.0.0.1"));

// Remove authorization from /auth-token
delete schema.paths["/auth-token"].post.security;
delete schema.paths["/auth-token/"].post.security;

// Remove cookie auth
delete schema.components.securitySchemes.cookieAuth;

await fs.writeFile(
  "./api-reference/openapi.json",
  JSON.stringify(schema, null)
);


// Keep iframe heights up to date
window.addEventListener("message", function (e) {
  const d = e.data;
  if (d.type !== "tako::resize") {
    return;
  }
  for (let iframe of document.querySelectorAll("iframe")) {
    if (iframe.contentWindow !== e.source) {
      continue;
    }
    iframe.style.height = d.height + "px";
  }
});


// Handle dark mode changes for visualizations
(function() {
  // Function to check dark mode and update all viz iframes
  function updateVizFrames() {
    const isDark = document.documentElement.classList.contains('dark');
    const vizFrames = document.querySelectorAll('iframe[data-is-viz="true"]');
    
    vizFrames.forEach(frame => {
      const baseSrc = frame.getAttribute('data-src');
      if (baseSrc) {
        frame.src = baseSrc + (isDark ? '?dark_mode=true' : '');
      }
    });
  }
  
  // Also update frames that may be added later
  function setupDynamicFrameHandler() {
    // Use MutationObserver to detect when new iframes are added
    const bodyObserver = new MutationObserver((mutations) => {
      mutations.forEach((mutation) => {
        if (mutation.addedNodes.length) {
          // Check if any new viz iframes were added
          const newFrames = document.querySelectorAll('iframe[data-is-viz="true"]:not([data-theme-handled="true"])');
          if (newFrames.length > 0) {
            updateVizFrames();
            // Mark frames as handled
            newFrames.forEach(frame => frame.setAttribute('data-theme-handled', 'true'));
          }
        }
      });
    });
    
    // Observe the entire document for added nodes
    bodyObserver.observe(document.body, { childList: true, subtree: true });
  }
  
  // Function to set up observers and initial state
  function initialize() {
    // Set up theme change detection - only looking at <html>
    const observer = new MutationObserver(() => {
      updateVizFrames();
    });
    
    // Start observing the HTML element's class attribute only
    observer.observe(document.documentElement, { 
      attributes: true,
      attributeFilter: ['class']
    });
    
    // Set initial state based on current dark mode
    updateVizFrames();
    
    // Set up handler for dynamically added frames
    setupDynamicFrameHandler();
    
    // Run updateVizFrames again after a short delay to catch any late-loading frames
    setTimeout(updateVizFrames, 500);
  }
  
  // Run immediately if DOM is already loaded
  if (document.readyState === 'loading') {
    // Still loading, so wait for the event
    document.addEventListener('DOMContentLoaded', initialize);
  } else {
    // DOM is ready, call initialize directly
    initialize();
  }
})();

Quickstart

Tako Docs

Overview

Tako is a search engine for visualizing and sharing the world's knowledge.

Introduction

Knowledge Cards

Data Sources

API Playground

Use natural language to search for Tako Knowledge Cards

Knowledge Search

Querying Tako

Grounding With Tako

Embedding Knowledge Cards

Here's a practical guide to using Tako with LlamaIndex to create a conversational AI agent:

LlamaIndex

This Python script sets up a basic MCP (Modular Conversational Platform) server integrated with Tako Knowledge Search API. The provided code creates an agent powered by MCP that uses Tako's Knowledge Search API to search for knowledge cards in response to user queries.

MCP Server

Create your MCP Server with Tako

Use Tako with MCP to create an agent that can answer questions about your data.

MCP with Notion

Integrate Tako MCP with Notion

Use Tako knowledge cards to ground your AI chatbot

Vercel AI SDK

Integrate Tako into your chatbot

Add interactive knowledge cards to your AI chatbot

Next.js AI Chatbot

Integrating Tako Cards into a Vercel Next.js AI Chatbot

Tako is an embeddable, composable knowledge graph sourced exclusively from authoritative structured data. Our API takes natural language queries or faceted searches as an input and returns iframe URLs that render responsive, interactive visualizations.

Getting Started

Authentication

Retrieve a list of collections. Collections are the primitive Tako uses to group data. They are useful for performing faceted searches.

Getting Started

Integrating Tako

Popular Integrations and Examples

Quickstart

Sign Up

Python SDK

Javascript SDK

API Reference