Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/JasonHonKL/spy-search/llms.txt

Use this file to discover all available pages before exploring further.

Spy Search is a self-hosted agentic search framework that orchestrates multiple AI agents — Planner, Searcher, and Reporter — to deliver fast, accurate answers and long-form research reports from live web content. It supports OpenAI, DeepSeek, Gemini, Grok, Anthropic, and Ollama out of the box, and costs nothing beyond your LLM API usage.

Quickstart

Get Spy Search running locally in under 5 minutes with a working configuration.

Docker Setup

Deploy the full stack — backend and frontend — with a single Docker Compose command.

Configuration

Learn how to configure your LLM provider, model, and agents via config.json.

API Reference

Explore all REST endpoints for search, reports, file management, and more.

What Spy Search Does

Spy Search routes your queries through a coordinated pipeline of AI agents. The Planner decomposes the task, the Searcher retrieves live web content via DuckDuckGo, and the Reporter synthesizes a coherent response or long-form report — all in one request.

Search Modes

Quick search, deep search, and academic arXiv search modes.

Report Generation

Generate ~2000-word research reports from live sources.

Local RAG

Query your own documents with ChromaDB vector search.

Get Started in 3 Steps

1

Clone and configure

Clone the repository and copy the example config to config.json. Set your LLM provider, model, and desired agents.
git clone https://github.com/JasonHonKL/spy-search.git
cd spy-search
cp config.example.json config.json
2

Run the setup script

Install Python dependencies and the Playwright browser engine.
python setup.py
3

Launch with Docker Compose

Start the backend API on port 8000 and the frontend on port 8080.
docker-compose up --build
Spy Search v0.3 is the current stable release. Check the Roadmap for planned features including MCP support, Google API integration, and multi-site search.

Build docs developers (and LLMs) love