deep-research README

The project's README from github.com/dzhng/deep-research. As one markdown file: README.md.

Open Deep Research

An AI-powered research assistant that performs iterative, deep research on any topic by combining search engines, web scraping, and large language models.

The goal of this repo is to provide the simplest implementation of a deep research agent - e.g. an agent that can refine its research direction over time and deep dive into a topic. Goal is to keep the repo size at <500 LoC so it is easy to understand and build on top of.

If you like this project, please consider starring it and giving me a follow on X/Twitter. This project is created by Duet.

How It Works

flowchart TB
    subgraph Input
        Q[User Query]
        B[Breadth Parameter]
        D[Depth Parameter]
    end

    DR[Deep Research] -->
    SQ[SERP Queries] -->
    PR[Process Results]

    subgraph Results[Results]
        direction TB
        NL((Learnings))
        ND((Directions))
    end

    PR --> NL
    PR --> ND

    DP{depth > 0?}

    RD["Next Direction:
    - Prior Goals
    - New Questions
    - Learnings"]

    MR[Markdown Report]

    %% Main Flow
    Q & B & D --> DR

    %% Results to Decision
    NL & ND --> DP

    %% Circular Flow
    DP -->|Yes| RD
    RD -->|New Context| DR

    %% Final Output
    DP -->|No| MR

    %% Styling
    classDef input fill:#7bed9f,stroke:#2ed573,color:black
    classDef process fill:#70a1ff,stroke:#1e90ff,color:black
    classDef recursive fill:#ffa502,stroke:#ff7f50,color:black
    classDef output fill:#ff4757,stroke:#ff6b81,color:black
    classDef results fill:#a8e6cf,stroke:#3b7a57,color:black

    class Q,B,D input
    class DR,SQ,PR process
    class DP,RD recursive
    class MR output
    class NL,ND results

Features

Requirements

Setup

Node.js

  1. Clone the repository
  2. Install dependencies:
npm install
  1. Set up environment variables in a .env.local file:
FIRECRAWL_KEY="your_firecrawl_key"
# If you want to use your self-hosted Firecrawl, add the following below:
# FIRECRAWL_BASE_URL="http://localhost:3002"

OPENAI_KEY="your_openai_key"

To use local LLM, comment out OPENAI_KEY and instead uncomment OPENAI_ENDPOINT and OPENAI_MODEL:

Docker

  1. Clone the repository

  2. Rename .env.example to .env.local and set your API keys

  3. Run docker build -f Dockerfile

  4. Run the Docker image:

docker compose up -d
  1. Execute npm run docker in the docker service:
docker exec -it deep-research npm run docker

Usage

Run the research assistant:

npm start

You'll be prompted to:

  1. Enter your research query
  2. Specify research breadth (recommended: 3-10, default: 4)
  3. Specify research depth (recommended: 1-5, default: 2)
  4. Answer follow-up questions to refine the research direction

The system will then:

  1. Generate and execute search queries
  2. Process and analyze search results
  3. Recursively explore deeper based on findings
  4. Generate a comprehensive markdown report

The final report will be saved as report.md or answer.md in your working directory, depending on which modes you selected.

Concurrency

If you have a paid version of Firecrawl or a local version, feel free to increase the ConcurrencyLimit by setting the CONCURRENCY_LIMIT environment variable so it runs faster.

If you have a free version, you may sometimes run into rate limit errors, you can reduce the limit to 1 (but it will run a lot slower).

DeepSeek R1

Deep research performs great on R1! We use Fireworks as the main provider for the R1 model. To use R1, simply set a Fireworks API key:

FIREWORKS_KEY="api_key"

The system will automatically switch over to use R1 instead of o3-mini when the key is detected.

Custom endpoints and models

There are 2 other optional env vars that lets you tweak the endpoint (for other OpenAI compatible APIs like OpenRouter or Gemini) as well as the model string.

OPENAI_ENDPOINT="custom_endpoint"
CUSTOM_MODEL="custom_model"

How It Works

  1. Initial Setup

    • Takes user query and research parameters (breadth & depth)
    • Generates follow-up questions to understand research needs better

  2. Deep Research Process

    • Generates multiple SERP queries based on research goals
    • Processes search results to extract key learnings
    • Generates follow-up research directions

  3. Recursive Exploration

    • If depth > 0, takes new research directions and continues exploration
    • Each iteration builds on previous learnings
    • Maintains context of research goals and findings

  4. Report Generation

    • Compiles all findings into a comprehensive markdown report
    • Includes all sources and references
    • Organizes information in a clear, readable format

Community implementations

Python: https://github.com/Finance-LLMs/deep-research-python

License

MIT License - feel free to use and modify as needed.

deep-research (dzhng/deep-research)

Searching Database…

19.8K ·GitHub
Deep-research
What is this?An AI-powered research assistant that digs into any topic iteratively: it combines search engines, web scraping, and large language models, refining its research direction as it goes.
Is this an app I run, or a library I add to my own project?A runnable TypeScript app, not a library: npm start drives a CLI research session, while npm run api boots an Express API server. The code is kept deliberately small so you can build on it.
How do I get it running?Clone the repo, run npm install, put your API keys in a .env.local file, then run npm start for the CLI or npm run api for the server. Node.js 22 is required.
What do I need in place before it can run?Node.js 22.x plus API keys: LLM calls go through the Vercel AI SDK with OpenAI and Fireworks adapters, and the web scraping needs a Firecrawl key.
Any rough edges I should know about before diving in?The author's stated goal is the simplest possible deep research agent, not a polished product — there is no test suite yet (the test script is a stub), so expect a lean codebase to extend.
Does it run on Windows and Mac, or is it Linux-only?It is a plain Node.js app, so it runs anywhere Node 22 does — macOS, Windows, or Linux.
Can I run it on my own server?Yes — the repo ships a Dockerfile and a docker-compose setup, so you can run it containerized and self-hosted on any server.
Does it need to be online the whole time?Yes — every step goes through cloud services: web searches, Firecrawl scraping, and LLM API calls, so there is no offline mode.
Can I freely use and modify it in my own projects?Yes — it is MIT-licensed, free to use and modify even commercially; just note the LLM and scraping APIs it calls may have their own costs.
What does the final output look like?A comprehensive markdown report that compiles all findings, includes all sources and references, and organizes everything in a clear, readable format.
How does it decide what to research next?It loops: each iteration takes new research directions based on previous learnings and maintains context of the research goals and findings, so it can deep dive into a topic over time.
Is there a Python version?Yes — the README lists community implementations, including a Python version of the agent, so you are not locked into TypeScript.
What are similar projects?gpt-crawler, canonry, Raven, wigolo, deer-flow, ChatDev, DeepAnalyze, trueforge
Which topics is it in?deep-research, search-engine, express, gpt, web-scraping, openai, research, agent, api, ai-agents, tool, llm
Creator
dzhng
Size
0.1 MB
Stars
19.8K
Updated
CommitApr 11
Age
1y
Commits · 90 Days