How to Run DeepSeek AI Locally on Android: A Step-by-Step Guide

Since its launch, ChatGPT has quickly become one of the top models. Google has been catching up with its Gemini model, and almost every company in 2025 has integrated AI into their apps and built services around it.

This also comes at a cost. OpenAI has been pouring billions of dollars into ChatGPT development; However, one Chinese AI company, with its DeepSeek model, which almost matches the capability of the ChatGPT top reasoning model, apparently has been developed at a fraction of the cost.

This has hit the US market and made us rethink the development of AI Chatbots and LLMs. DeepSeek is so disruptive that now you can run LLM reasoning models known for their thinking capabilities before answering your questions locally on your device. We already have a detailed guide on how to run DeepSeek locally on your Windows and Mac. In this guide, We will look at installing and running the DeepSeek R1 model locally on your Android smartphone.

Minimum Device requirements to install and run DeepSeek locally on your device

Since LLM requires resources to function, there are set minimum system requirements for installing and using DeepSeek R1 locally on your device. Remember, these are system requirements for the smaller 1.5B model. A 7,8,32B, 70 billion model also requires higher system requirements.

With that in mind, if you want an Android flagship smartphone with the system requirements, let’s go to the step-by-step guide and install and run DeepSeek AI on your device.

Step-by-Step Guide to run DeepSeek locally on your Device

You can install and run Deep Seek AI locally on your device in several ways. In this post, you will learn how to use Termun to install and run local AI models locally. However, this requires working with the terminal; if you want an easy method, you can use the Pocketpal AI app. See the second method.

Step 1: Install Termux

Step 1.1: Download the Termux app from the official GitHub page.

Termux is a terminal emulator Android app that brings its command line to mobile devices like Android. It lets you execute command-line programs, install Linux packages, execute shell commands, and allows you to install tools like Python and SSH on your device. Most importantly, unlike many apps, it does not require root access, which means anyone can install this app without rooting or jailbreaking the device, making it safe and easy to use for businesses.

The app is not available on Google Play Store or App Store after it was removed from the Play Store. However, you can easily install it from third-party sources and websites like Github. We also have a detailed guide on official and safe links for Termux. You can visit our guide and install the latest APK.

For this guide, you can install Termux from the GitHub page.

Step 1.2: Visit Github Termux Download Page

Step 1.3: Scroll down, and under the F-Driod section, click on the download link (here), or you can use this direct download link to visit the F-Drop page.

Step 1.4: Scroll down and find the latest version of the app (as of writing this article, the latest version is 0.119.0-beta.1 (1020); it worked fine for me; if you are having issues, you can download the stable version. Scroll down and install Version 0.118.1 (1000) version).

Step 1.5: Click on the Download APK file to download.

Step 1.6: Go to the downloaded file and click on the install. If you need permissions, enable ‘Install from Unknown Sources’ in the Android settings. Install and open Termux on your device.

Step 1.7: If there is a security prompt, click on the More option and tap the Install anyway option, and the app will be successfully installed on your device.

Step 2: Now let’s Set Up Termux

Once there, you must set up Termux to run on your device.

To do this, open the app and grant storage permission. By default, Termux doesn’t have permission to access your phone files after installing the app. Using this command, you can grant permission to access Android files, read and store data, and use file management commands.

Step 2.1: Open the Termux app on your Android smartphone, enter this command, and press enter. termux-change-repo

However, while doing this, you might encounter common errors such as the ‘Permission Denied’ error, which occurs when Android denies granting permission. Go to Settings > App Permissions to fix this and manually enable storage permissions.

Another common error users face is the ‘No such file or directory‘ error, meaning the storage is not initialized. In this case, close the Termux app, reopen it, force stop it, and reopen it after running the command.

Step 2.2: Next, you need to update the package repository. You can use the command below to update the package repository. In the Termux app, Enter this command: apt update && apt upgrade -y

You must update the Termux repository to update all the packages. This will help you get the latest security patches and prevent compatibility issues. However, when doing this, ensure you have an active internet connection to fetch the latest packages from the server and install them on your Termux app.

Step 3. Next, Install PRoot Distro & Debian 12

Next, you have to install Proot-Distro and Debian within Termux. The Proot package allows you to run Linux distributions like Debian or Ubuntu inside Termux.

Step 3.1: To Install PRoot Distro, Enter this command: pkg install proot-distro -y in the Termux app and press enter.

While doing this, encountering a ‘Package not found’ error means Termux has not been updated. Again, run apt update && apt upgrade -y command, and also ensure you have an active internet connection while using this command.

Next, you need to install Debian in Termux. Debian is a free and open-source Linux distribution that is lightweight and runs on your Android using Proot-Distro. You can install the latest version of Debian using the following command below. Also, ensure you have enough storage on your device and a stable internet connection.

Step 3.2: Enter this command in the terminal: proot-distro install debian

Step 3.3:Next, you must log in to Debian using the following command: proot-distro login Debian.

Step 4. Install TMUX & Ollama To Install and Use DeepSeek R1

Next, you must install TMUX and Ollama on your device to run DeepSeek R1 locally. To do this, you first need to update the Debian package list. You can use the following command to update all your Debian package lists to the latest versions.

Step 4.1: Enter the following command: apt update && apt upgrade -y to update the Debian package list.

Next, you need to install TMUX to manage multiple terminal sessions. TMUX allows you to run multiple terminal sessions inside a single window. The primary reason for installing TMUX is that it helps you keep your AI model running in the background, even if you accidentally disconnect. It also comes with other benefits, such as splitting your terminal into multiple panes for better multitasking and running multiple tasks simultaneously.

Step 4.2: Use the following command: apt install tmux -y to install TMUX

Next, you need to install Ollama, a tool designed for Android that lets you run AI models locally on your device. You can run DeepSeek R1 and Meta Llama locally on your device using this tool. It also supports offline AI interfaces like ChatGPT.

Step 4.3: Once you have done all this, you must install and run DeepSeek R1 locally on your device. Use the following command to install Ollama: curl -fsSL https://ollama.ai/install.sh | sh

Step 5. Install & Run DeepSeek R1

Next, you need to start the Ollama server inside Termux. To do this, use the following command:

Step 5.1: Step Start Ollama server inside TMUX: tmux ollama serve

Step 5.2: Open a new TMUX window and split the terminal into multiple panes. This can be helpful when running the AI model in one pane, monitoring logs, or entering commands in another terminal. Use the following shortcut to split the terminals: Click on the Control+B+”

Next, you must install the DeepSeek R1 1.5B model locally on your device. There are several models for DeepSeek R1 models. Since Android has limited resources and running these models locally requires a lot of resources, the 1.5B model is the most reliable model for most Android devices. Here is a list of models that DeepSeek R1 supports.

Download your device’s DeepSeek R1 1.5B local model using the following command. This downloads and saves it on your device. Ensure your device has a minimum of required storage for each model.

Step 5.3: In the new Termux windows, use the command to install DeepSeek R1 (1.5B model): ollama pull deepseek-r1:1.5b

You can install the 8B model using the same command if you have a high-end flagship device. Similarly, if you have other high-end devices, you can install the 15B parameter model locally on your device.

For your information, the more model parameters, the more capable the model is. For most use cases, the 1.5B parameter model is an easy pick for just conversations. However, the 8B parameter model is reasonable and often recommended if you want more knowledge and better performance.

Thats it. Now, you have successfully installed DeepSeek R1 on your device. Now that DeepSeek R1 is installed, you can use Termux to run it locally on your device.

Step 6. Run DeepSeek Locally on Your Device

Step 6.1: To run DeepSeek R1 in the terminal, use the command: ollama run deepseek-r1:1.5b.

Make sure that you use the correct model number in the command. If you installed the 1.5B model, use the same model name in the command. If you installed the 8B model, use the command.

Now, you can try interacting with the model using text input. Since everything runs through the terminal, there will be no GUI (Graphical User Interface) to interact with the ChatGPT; you can use the text input and get results in the same terminal.

Depending on your device’s local storage and processing power, the model will generate output based on available resources. Here are the most common Termux shortcuts for managing your AI sessions locally on your device. You can use CTRL + C to stop the AI model, CTRL + D to exit, return to the model, and more.

Use PocketPal to Run DeepSeek R1 Locally on Your Device

If you find the above method overwhelming, you can use PocketPal AI, a free app that installs and runs AI models locally on your device. Follow the steps below.

Step 1: Install the PocketPal app from the Google Play Store. 

Step 2: Open the app and go to models

Step 3: Click the plus icon and tap the Add from Hugging Face.

Step 4: Use the search bar in the button and search for DeepSeek R1 models. 

Step 5: Select the model based on your device storage limit and click the download icon. Wait for the model to download.

Step 6: Once downloaded, Return to the home screen, select the model, and click on the Load.

This will load DeepSeek R1 locally on your device. Unlike the above method, it gives you a chat graphical interface to interact with the model. You can also turn off the internet connection if you need to.

Run the AI Model Locally on Your Mobile

Even though DeepSeek is disruptive and very beneficial to the consumer, several reports claim that DeepSeek can send data to Chinese companies. With this model running on your device locally, the data stays on your device. Plus, if you have good specifications for a flagship smartphone, the DeepSeek is faster and has no slow loading times to respond, which is generally the case with the cloud models. All this makes DeepSeek R1 a great resource for installing, running, and using LLM locally on your device.

I have shared your detailed, comprehensive step-by-step help. If you still encounter any issues or errors while installing, you can comment below, and I will try to help you. I hope you find this guide helpful. Also, check out our guide on installing DeepSeek locally on your Mac and Windows; that would be very helpful for your professional and student work tasks.