🚀 agenticSeek: Local AI Assistant Powered by DeepSeek Agents

A fully local AI assistant using a swarm of DeepSeek agents, capable of:
✅ Code execution (Python, Bash)
✅ Self-correcting code execution
✅ Routing system, select the best agent for the task ✅ Speech-to-text & text-to-speech
✅ Web browsing (under development, see dev)

🛠️ Work in Progress – Looking for contributors! 🚀

---

🌟 Why?

• Privacy-first: Runs 100% locally – no data leaves your machine
• ️ Voice-enabled: Speak and interact naturally
• Self-correcting: Automatically fixes its own code
• Multi-agent: Use a swarm of agents to answer complex questions
• Web browsing (not implemented yet): Browse the web and search the internet
• Knowledge base (not implemented yet): Use a knowledge base to answer questions

---

Installation

1️⃣ Install Dependencies

Make sure you have Ollama installed, then run:

pip3 install -r requirements.txt

2️⃣ Download Models

Download the deepseek-r1:7b model from DeepSeek

ollama pull deepseek-r1:7b

3️⃣ Run the Assistant (Ollama)

Start the ollama server

ollama serve

Change the config.ini file to set the provider_name to ollama and provider_model to deepseek-r1:7b

[MAIN]
is_local = True
provider_name = ollama
provider_model = deepseek-r1:7b

Run the assistant:

python3 main.py

4️⃣ Alternative: Run the Assistant (Own Server)

On the other machine that will run the model execute the script in stream_llm.py

python3 stream_llm.py

Get the ip address of the machine that will run the model

ip a | grep "inet " | grep -v 127.0.0.1 | awk '{print $2}' | cut -d/ -f1

Change the config.ini file to set the provider_name to server and provider_model to deepseek-r1:7b. Set the provider_server_address to the ip address of the machine that will run the model.

[MAIN]
is_local = False
provider_name = server
provider_model = deepseek-r1:14b
provider_server_address = x.x.x.x:5000

Run the assistant:

python3 main.py

Current capabilities

• All running locally
• Reasoning with deepseek R1
• Code execution capabilities (Python, Golang, C)
• Shell control capabilities in bash
• Will try to fix errors by itself
• Routing system, select the best agent for the task
• Fast text-to-speech using kokoro.
• Speech-to-text using distil-whisper/distil-medium.en
• Memory compression (reduce history as interaction progresses using summary model)
• Recovery: recover last session from memory

UNDER DEVELOPMENT

• Web browsing
• Knowledge base RAG