Good day! 🫡 🥷

Today we’re gonna have a look on a Open Source Deep-Researcher alternative to the costly $200 OpenAI version. You’ll have the full step-by-step tutorial for how you can run it locally.

But first, a funny meme:

Let's have a look at today's menu ↓

Today we're diving into 🤿

• Open Deep-Research - What is it? What can I use it for? 🤔 
• Open Deep-Research - Full guide to set it up 💻️ 
• Prompt Share - Flux 🎨 
• Fresh AI News 🗞️ 
• New Trending AI tools 🧰 

Let's go! ↓

_{Together with Synthflow :}

🎭 Samurai Spotlight - Synthflow

Create, Publish & Earn with Synthflow AI Voice Agents Marketplace

• Discover templates for routine/repetitive tasks like lead qualification and managing appointments.

• Publish your own Voice AI solutions to help businesses thrive—and earn commissions.

• Access custom actions that automate CRM updates, appointment scheduling, and more.

Build Your AI Agent Now

🤔 What is ‘Open Deep-Research’ and what can you use it for:

First of all, what is Open Deep-Research??

Open Deep Research is a framework that lets an AI “agent” search for, collect, and summarize information across the web (and other sources). You can compare it to putting a digital researcher on autopilot!

In simple terms, it can plan out your search, navigate websites or APIs, fetch the relevant details, skillfully synthesize them, and present you with well-organized results—much like a dedicated research assistant would.

You can carefully watch and analyze each step the AI is taking, and you can refine or redirect its process at any stage.

A few examples of what it can be used for:

1. Gathering Data from Multiple Sources:

The AI agent browses different sites on a topic of your choice, collecting key information (like answers, abstracts, or summaries), and organizes it all in one place in e.g. a report or Google Docs.

2. Complex Question Answering:

Instead of just looking at the first result from a search engine, the agent can combine multiple sources, cross-reference data, and craft comprehensive answers to your queries.

3. Collaborative Research:

Work on big research projects (like writing a report or compiling facts for an article) by letting the AI dig up the data, while you focus on thinking through and interpreting the results.

4. Academic Literature Reviews:

The AI can look up papers on Google Scholar (or open-source archives), gather the key points, highlight relevant sections, and produce a quick literature review.

5. Market Analysis:

The AI can gather data from news sites, industry blogs, or social media to gauge trends and competition for a business idea or product you have in mind.

6. Tech Troubleshooting:

The AI can quickly comb through forums, official docs, or Q&A platforms to find detailed, step-by-step solutions to your technical issues.

7. Tracking Public Opinions:

The AI can check multiple social platforms and compile user sentiments about your company or a public figure, categorizing sentiments as positive, negative, or neutral.

So, ready to dive in? In the next section, I’ll show you how to get set up—and it only takes a few minutes!

💻️ How to Set Up Open Deep Research for FREE
FULL step-by-step Guide:

Step 1: Install Python

Open Deep Research is built in Python. Make sure you have Python 3.9 or later installed.

Download Python and Git:

• Visit the official Python website.
• Visit the git-scm.com/download/win
• Choose the version for your operating system (Windows, macOS, or Linux).

Install Python and Git:

• Windows: Run the installer, check the box “Add Python to PATH”.
• macOS/Linux: Follow the instructions provided on the site.

Verify Installation:

• Open a terminal (on Windows, hit Win + R, type cmd, press Enter).
• Run:

python --version
git --version

Make sure the response shows Python 3.9+.

Step 2: Create a Project Folder and Virtual Environment

It’s good practice to set up a dedicated virtual environment for managing dependencies.

1. Create a new folder (e.g. OpenDeepResearch) wherever you like (Desktop, Documents, etc.).

2. Open your terminal in that folder. For instance, on Windows:

cd C:\Users\YOUR_USER\Desktop\OpenDeepResearch

3. Create and activate the virtual environment:

python -m venv env

Activate the environment:

• Windows:

  env\Scripts\activate
  

• macOS/Linux:

  source env/bin/activate
  

You should see (env) in your terminal prompt indicating that you’re inside the virtual environment.

Step 3: Clone the smolagents Repository and Install Requirements

Open Deep Research is part of Hugging Face’s smolagents repository.

1. Clone the repo from GitHub:

git clone https://github.com/huggingface/smolagents.git

2. Go into the repo folder:

cd smolagents

3. Install the top-level smolagents package:

pip install -e .

4. Move to the open_deep_research example folder:

cd examples/open_deep_research

5. Install the example’s specific dependencies:

pip install -r requirements.txt

6. Install smolagents[litellm]:

pip install smolagents[litellm]

If you encounter installation errors, verify your Python and Git versions and your system dependencies.

Step 4: Set Up Your Model/LLM Credentials

By default, this example can use OpenAI’s API. If you’re using a different provider or a local model, adjust accordingly, but here’s how to set it up with OpenAI API.

1. Obtain Your OpenAI API Key:
   • Sign up at OpenAI’s API page.
   • Go to your account’s API Keys section.
   • Click “Create new secret key” and copy it.

2. Set the OPENAI_API_KEY in your environment or a .env file:

   # Windows:
   SET OPENAI_API_KEY=sk-YourKeyHere
   
   # macOS/Linux:
   export OPENAI_API_KEY=sk-YourKeyHere
   

   Or create a .env file in your open_deep_research folder with:

   OPENAI_API_KEY=sk-YourKeyHere

   Find or set up your OpenAI API key from here: https://platform.openai.com/api-keys

Step 5: Run a Basic Example

Inside the open_deep_research folder, there’s a run.py script that showcases how the code-based agent works.

1. Open your terminal in the open_deep_research folder.

2. Run:

python run.py "How many studio albums did Mercedes Sosa release before 2007?"

3. Set up and copy-paste your Huggingface token, guide here: https://huggingface.co/docs/hub/en/security-tokens

If all is well, the agent will start running locally. It uses a “code-based” approach for chaining actions (like browsing text-based pages, reading files, etc.).

Examples & Next Steps
• Try different prompts or tasks. Look at run.py to see how to adjust the agent’s settings.
• Experiment with an open-source model (like “DeepSeek R1” mentioned in the blog) by modifying the model call inside run.py.
• If you want more advanced browsers or tools, you can contribute or install other tools and adjust them in this agent.

That’s about it. Pretty simple. Please let me know if you run into any issues :)

🎯 Prompt Share:

I love generating AI images, so below I'll share some prompts with you:

📰 Latest AI News

🧰 Tools of the Week

1. ChatLLM by Abacus

Access to all the biggest LLMs.

Artifacts directly in browser, just like with Claude.

Build complex AI Agents.

Save me $10 each month compared to other LLMs.

2. Gamma AI

AI-powered presentation and document creator.

Transform text into stunning slides and webpages instantly.

No design skills needed, with built-in collaboration features.

3. Otio AI

Professional AI researcher.

This is the fastest way to do research

Perfect for students and researchers.

Thank you for reading!

What's on your mind?

Feel free to share your ideas with me at: _{[email protected]} 

If your idea is good, I’ll make sure to bring it to life. Just send them to the above e-mail and I’ll get to work on making them a reality.