Nutlope/turboseek

An AI search engine inspired by Perplexity

What it solves

TurboSeek is an open-source alternative to commercial AI search engines like Perplexity. It lets you ask a question in plain language and get a streamed, AI-written answer that is grounded in live web search results, rather than relying on a static knowledge base.

How it works

When a user submits a question, TurboSeek runs through a five-step pipeline:

  1. It sends the question to a search API (described as the Bing Search API in the steps, while Exa.ai is listed in the tech stack).
  2. It takes the top 6 results from that search and scrapes the text from each link.
  3. That scraped text becomes the context.
  4. It sends the user's question plus the context to an OpenAI gpt-oss model and streams the answer back.
  5. It makes a separate request to Llama 3.1 8B to generate 3 related follow-up questions the user can ask.

The app is built with Next.js (App Router) and Tailwind for the front end, uses Together AI for model inference, and includes Helicone for observability and Plausible for analytics.

Who it's for

TurboSeek is for developers who want to run their own AI search engine, or who want to learn how to build one by following the linked tutorial. It is also useful for anyone who wants an open, self-hosted way to get AI answers from live web content.

Highlights

  • Open source and self-hostable
  • Streams answers in real time
  • Automatically suggests follow-up questions
  • Uses a simple, familiar stack: Next.js, Tailwind, Together AI, and a search API
  • Includes a step-by-step tutorial to build it yourself

Related

  • Project
    JShollaj/awesome-llm-interpretability

    A curated collection of tools, research papers, and articles focused on understanding and interpreting the internal workings of Large Language Models.

  • Project
    Jeffser/Alpaca

    A beginner-friendly Ollama client that allows users to manage and chat with local AI models and cloud-hosted providers in a single interface.

  • Project
    zarazhangrui/youtube-digest

    A Chrome extension that turns YouTube videos into learning resources by providing AI-generated summaries, bilingual transcripts, and timestamped notes using the user's own API keys.

  • Project
    lizhe2004/Awesome-LLM-RAG-Application

    A curated, comprehensive collection of resources, academic papers, and open-source tools for building and optimizing Retrieval-Augmented Generation (RAG) applications.