{"id":4788,"date":"2025-06-23T07:04:01","date_gmt":"2025-06-23T07:04:01","guid":{"rendered":"https:\/\/mailitics.com\/index.php\/2025\/06\/23\/project_i_just_opensourced_a_plugin_to_stop_ai\/"},"modified":"2025-06-23T07:04:01","modified_gmt":"2025-06-23T07:04:01","slug":"project_i_just_opensourced_a_plugin_to_stop_ai","status":"publish","type":"post","link":"https:\/\/mailitics.com\/index.php\/2025\/06\/23\/project_i_just_opensourced_a_plugin_to_stop_ai\/","title":{"rendered":"[Project] I just open-sourced a plugin to stop AI from hallucinating your schemas"},"content":{"rendered":"<p>    [Project] I just open-sourced a plugin to stop AI from hallucinating your schemas<br \/>\n \t<BR><br \/>\n<BR><\/BR><br \/>\n    <!-- no image --><br \/>\n \t<BR><br \/>\n<BR><\/BR><\/p>\n<div>\n<!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>Hey <a href=\"https:\/\/www.reddit.com\/r\/datascience\">r\/datascience<\/a> \ud83d\udc4b<\/p>\n<p>Using AI tools like Copilot or Cursor can be a total headache for data science work. You&#8217;re trying to join tables, and it confidently suggests <code>customer_id<\/code> when your table <em>actually<\/em> uses <code>cust_pk<\/code>. Or worse, it just invents tables that don&#8217;t even exist. Sound familiar?<\/p>\n<p>The problem is, these AI assistants are blind to your database schemas. They&#8217;re great for general code, but for data science, they constantly hallucinate table names, column structures, and relationships. It turns a supposed productivity boost into an endless game of whack-a-mole.<\/p>\n<p>I got so fed up copy-pasting schemas into ChatGPT, I decided to build <a href=\"https:\/\/github.com\/kruskal-labs\/toolfront\">ToolFront<\/a>. It&#8217;s a free, open-source IDE plugin that finally gives your AI assistant a smart, safe way to <strong>understand all your databases and query them<\/strong>.<\/p>\n<h1>So, what does it do?<\/h1>\n<p>ToolFront equips your coding AI (Cursor\/Copilot\/Claude) with a set of <strong>read-only database tools<\/strong>:<\/p>\n<ul>\n<li>\n<code>discover<\/code>: See all your connected databases.<\/li>\n<li>\n<code>scan<\/code>: Find tables by name or description.<\/li>\n<li>\n<code>inspect<\/code>: Get the exact schema for any table \u2013 no more guessing!<\/li>\n<li>\n<code>sample<\/code>: Grab a few rows to quickly see the data.<\/li>\n<li>\n<code>query<\/code>: Run read-only SQL queries directly.<\/li>\n<li>\n<code>learn<\/code> <strong>(The Best Part)<\/strong>: Finds the most relevant historical queries written by you or your team to answer new questions. Your AI can actually learn from your team&#8217;s past SQL!<\/li>\n<\/ul>\n<h1>Connects to what you&#8217;re already using<\/h1>\n<p>ToolFront supports the databases you&#8217;re probably already working with:<\/p>\n<ul>\n<li>\n<strong>Snowflake<\/strong>, <strong>BigQuery<\/strong>, <strong>Databricks<\/strong>\n<\/li>\n<li>\n<strong>PostgreSQL<\/strong>, <strong>MySQL<\/strong>, <strong>SQL Server<\/strong>, <strong>SQLite<\/strong>\n<\/li>\n<li>\n<strong>DuckDB<\/strong> (Yup, analyze local CSV, Parquet, JSON, XLSX files directly!)<\/li>\n<\/ul>\n<h1>Why you&#8217;ll love it<\/h1>\n<ul>\n<li>\n<strong>Faster EDA<\/strong>: Explore new datasets without constantly jumping to docs.<\/li>\n<li>\n<strong>Easier Onboarding<\/strong>: Get new team members productive on complex data warehouses quicker.<\/li>\n<li>\n<strong>Smarter Ad-Hoc Analysis<\/strong>: Get AI help without context-switching.<\/li>\n<\/ul>\n<p>If you&#8217;re a data scientist who uses AI assistants, I genuinely think ToolFront can make your life a lot easier.<\/p>\n<p>I&#8217;d love your feedback, especially on what database features are most crucial for your daily work.<\/p>\n<p><strong>GitHub Repo<\/strong>: <a href=\"https:\/\/github.com\/kruskal-labs\/toolfront\">https:\/\/github.com\/kruskal-labs\/toolfront<\/a><\/p>\n<p>A \u2b50 on GitHub really helps with visibility!<\/p>\n<\/p><\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/Durovilla\"> \/u\/Durovilla <\/a> <br \/> <span><a href=\"https:\/\/www.reddit.com\/r\/datascience\/comments\/1li6bww\/project_i_just_opensourced_a_plugin_to_stop_ai\/\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datascience\/comments\/1li6bww\/project_i_just_opensourced_a_plugin_to_stop_ai\/\">[comments]<\/a><\/span>\n<\/div>\n<p> \t<BR><br \/>\n <BR><\/BR><br \/>\n    \/u\/Durovilla<br \/>\n \t<BR><br \/>\n<BR><\/BR><br \/>\n<a href=\"https:\/\/www.reddit.com\/r\/datascience\/comments\/1li6bww\/project_i_just_opensourced_a_plugin_to_stop_ai\/\">Go to original source<\/a><br \/>\n \t<BR><br \/>\n <BR><\/BR><\/p>\n","protected":false},"excerpt":{"rendered":"<p>[Project] I just open-sourced a plugin to stop AI from hallucinating your schemas Hey r\/datascience \ud83d\udc4b Using AI tools like Copilot or Cursor can be a total headache for data science work. You&#8217;re trying to join tables, and it confidently suggests customer_id when your table actually uses cust_pk. Or worse, it just invents tables that [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[62,99],"tags":[98,84,163],"class_list":["post-4788","post","type-post","status-publish","format-standard","hentry","category-aimldsaimlds","category-datascience","tag-ai","tag-data","tag-your"],"_links":{"self":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/4788"}],"collection":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/comments?post=4788"}],"version-history":[{"count":0,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/4788\/revisions"}],"wp:attachment":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/media?parent=4788"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/categories?post=4788"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/tags?post=4788"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}