Skip to content

Tater is a local-native AI assistant powered by Cerberus — a structured planning engine. It runs on local LLMs through OpenAI-compatible endpoints and uses modular tools to plan, act, and evolve.

License

Notifications You must be signed in to change notification settings

TaterTotterson/Tater

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

689 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tater AI Assistant

Tater is an AI assistant that connects to any OpenAI-compatible LLM, giving you a powerful set of AI-driven tools. It includes a WebUI for setup and private chats, and works across Discord, Telegram, IRC, Matrix, Home Assistant, HomeKit, and even the OG Xbox via XBMC4Xbox

Main website: taterassistant.com


Cerberus

Tater runs on Cerberus — a closed-loop Planner → Doer → Checker architecture designed for reliable tool execution.


🌐 Tater Platform Overview

Platform Description
discord Full-featured Discord bot that runs compatible Verba Plugins and supports rich interactions, media output, and background jobs.
homeassistant Native integration for Home Assistant, allowing Tater to act as a voice-enabled AI assistant through the Assist pipeline and control smart devices directly.
ha_automations Lightweight Home Assistant automation-only endpoint for direct tool execution. Designed for fast, reliable automations like camera events, doorbell alerts, weather summaries, and dashboard sensors. Intended to be used with the Tater Automations custom component, which provides a native “Call Tater automation tool” action in Home Assistant.
ai_task Built-in scheduled task runner for timed and recurring AI jobs, with delivery routed through notifier platforms (Discord, Telegram, Matrix, IRC, Home Assistant, and more).
memory_platform Background memory extraction platform that incrementally scans chat history, builds structured user/room memory profiles in Redis, and injects compact memory context into Cerberus.
homekit Siri / Apple Shortcuts integration for HomeKit, enabling “Hey Siri, ask Tater…” voice commands, per-device conversation sessions, and Verba Plugin-backed actions.
irc Lightweight IRC bot that responds to mentions and runs compatible Verba Plugins.
matrix Modern Matrix client with end-to-end encryption support, Markdown rendering, and full Verba Plugin compatibility — bringing Tater to federated chat networks like Element and Cinny.
telegram Telegram bot integration with chat allowlists, DM user restrictions, queued notifications, media delivery, and Verba Plugin-backed tool execution.
xbmc OG Xbox integration for XBMC4Xbox, provided by the custom Cortana-powered skin and script at skin.cortana.tater-xbmc, enabling an on-console AI assistant with chat history, quick asks, and Verba Plugin-powered actions.
webui Streamlit-based WebUI for chatting, configuring Verba Plugins, and managing settings.

🧩 Tater Verba Plugin Ecosystem

Tater now uses a remote Verba Plugin Store.
Verba Plugins are no longer bundled with Tater — they are installed, updated, and restored automatically from the Tater Shop.

🛒 Tater Verba Plugin Store

All Verba Plugins, versions, descriptions, and update history now live here:

👉 https://github.com/TaterTotterson/Tater_Shop?target=https://github.com


🔍 Browsing Verba Plugins

From the WebUI you can:

  • Search by name or description
  • Filter by platform (Discord, WebUI, Home Assistant, etc.)
  • See installed vs store versions
  • One-click install, update, remove
  • Bulk “Update All”

♻️ Auto-Restore

When Tater starts:

  • Any Verba Plugin that was enabled in Redis but missing on disk
  • is automatically re-downloaded from the store.

No config loss. No manual installs. No volume mapping required.

Note:

  • Tater currently recommends using qwen3-coder-next, qwen3-next-80b, gpt-oss-120b, qwen3-coder-30b or Gemma3-27b

Installation

Note:

  • Tater currently recommends using qwen3-coder-next, qwen3-next-80b, gpt-oss-120b, qwen3-coder-30b or Gemma3-27b
unraid_logo_black-339076895

Tater is available in the Unraid Community Apps store.

You can install both:

  • Tater
  • Redis Stack

directly from the Unraid App Store with a one-click template.

Important for Docker/Unraid persistence:

  • Add a path mapping for /app/agent_lab (container) -> /mnt/user/appdata/tater/agent_lab (host example).
  • Without this mapping, data in /agent_lab (logs/downloads/documents/workspace) can be lost on container rebuilds/updates.

🏠 Home Assistant

A dedicated Home Assistant add-on repository is available here:

https://github.com/TaterTotterson/hassio-addons-tater?target=https://github.com

Add the Tater add-on repository

Click the button below to add the repository to Home Assistant:

Add Repository to Home Assistant

Once added, the following add-ons will appear in the Home Assistant Add-on Store:

  • Redis Stack – required for Tater memory, Verba Plugins, and automations
  • Tater AI Assistant – the main Tater service

Install order

  1. Install and start Redis Stack
  2. Install Tater AI Assistant
  3. Configure your LLM and Redis settings in the Tater add-on
  4. Start Tater

This is the recommended setup for most users and provides the smoothest experience.

🔌 Home Assistant Integration (Optional Enhancements)

Tater AI supports deeper integration with Home Assistant through a couple of companion repositories. These extend Tater’s usefulness in smart home setups by enabling voice/text conversation control and native automation triggers.

📍 Tater-HomeAssistant

https://github.com/TaterTotterson/Tater-HomeAssistant?target=https://github.com

A Home Assistant custom integration that allows Tater to function as a Conversation Agent inside Home Assistant’s Assist pipeline. This enables voice or text interactions from Home Assistant to be routed directly to your Tater backend, where Verba Plugins can be executed and contextual responses returned.

Key benefits:

  • Use Tater as a native voice and text assistant within Home Assistant
  • Routes Assist queries directly to your running Tater AI backend
  • Supports Verba Plugins that implement handle_homeassistant(...)
  • Maintains conversation context for more natural, multi-turn interactions

This integration is required if you want Tater to participate directly in Home Assistant conversations or voice control.

⚙️ tater_automations

https://github.com/TaterTotterson/tater_automations?target=https://github.com

A Home Assistant automation-focused custom component that exposes Tater’s tools as native Home Assistant automation actions. This allows Home Assistant automations to call specific Tater tools directly, without REST calls, scripts, or YAML workarounds.

Key benefits:

  • Adds a native “Call Tater automation tool” action in Home Assistant automations
  • Designed for fast, reliable, automation-only execution
  • Ideal for camera events, alerts, summaries, and AI-driven logic
  • Integrates cleanly into Home Assistant’s automation editor and UI

This component is required if you want to trigger Tater tools directly from Home Assistant automations.


Local Installation (Advanced)

Prerequisites

  • Python 3.11
  • Redis-Stack
  • OpenAI API–compatible LLM app (such as Ollama, LocalAI, LM Studio, Lemonade, or OpenAI API)
  • Docker (optional, for containerized deployment)

Setting Up Locally

  1. Clone the Repository
git clone https://github.com/TaterTotterson/Tater.git
  1. Navigate to the Project Directory
cd Tater

Note:

  • Run Tater inside a Python virtual environment so dependencies stay isolated and easy to manage (recommended). Quickstart:
    python -m venv .venv
    source .venv/bin/activate
  1. Install Dependencies

Using pip, run:

pip install -r requirements.txt
  1. Configure Environment Variables

Create a .env file in the root directory.
Below are example configurations for local LLM backends (Ollama, LM Studio, LocalAI) and ChatGPT (GPT-4o, etc.).


Example: Local backend (Ollama, LM Studio, LocalAI)

LLM_HOST=127.0.0.1  
LLM_PORT=11434  
LLM_MODEL=gemma3-27b-abliterated  
REDIS_HOST=127.0.0.1  
REDIS_PORT=6379  

Example: ChatGPT (GPT-4o, etc.)

LLM_HOST=https://api.openai.com  
LLM_PORT=  
LLM_MODEL=gpt-4o  
LLM_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxx  
REDIS_HOST=127.0.0.1  
REDIS_PORT=6379  

Note: When using ChatGPT, leave LLM_PORT blank.
Tater will automatically connect using HTTPS without appending a port number.

  1. Run the Web UI

Launch the web UI using Streamlit:

streamlit run webui.py

Docker

1. Pull the Image

Pull the prebuilt image with the following command:

docker pull ghcr.io/tatertotterson/tater:latest

2. Configuring Environment Variables

Ensure you supply the required environment variables. You can pass these using the -e flag when starting the container.


Example: Local backend (Ollama, LM Studio, LocalAI)

docker run -d --name tater_webui \
  -p 8501:8501 \
  -p 8787:8787 \
  -p 8788:8788 \
  -p 8789:8789 \
  -p 8790:8790 \
  -e TZ=America/Chicago \
  -v /etc/localtime:/etc/localtime:ro \
  -v /etc/timezone:/etc/timezone:ro \
  -e LLM_HOST=127.0.0.1 \
  -e LLM_PORT=11434 \
  -e LLM_MODEL=gemma3-27b-abliterated \
  -e REDIS_HOST=127.0.0.1 \
  -e REDIS_PORT=6379 \
  -v /agent_lab:/app/agent_lab \
  ghcr.io/tatertotterson/tater:latest

Example: ChatGPT (GPT-4o, etc.)

docker run -d --name tater_webui \
  -p 8501:8501 \
  -p 8787:8787 \
  -p 8788:8788 \
  -p 8789:8789 \
  -p 8790:8790 \
  -e TZ=America/Chicago \
  -v /etc/localtime:/etc/localtime:ro \
  -v /etc/timezone:/etc/timezone:ro \
  -e LLM_HOST=https://api.openai.com \
  -e LLM_PORT= \
  -e LLM_MODEL=gpt-4o \
  -e LLM_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxx \
  -e REDIS_HOST=127.0.0.1 \
  -e REDIS_PORT=6379 \
  -v /agent_lab:/app/agent_lab \
  ghcr.io/tatertotterson/tater:latest

Note: When using ChatGPT, leave LLM_PORT blank.
Tater will automatically connect using HTTPS without appending a port number.

Tip: The runtime data lives in /app/agent_lab inside the container.
If you don’t mount it to the host, /agent_lab data can be lost when the container is rebuilt or updated.

Unraid note: add a container path mapping for /app/agent_lab to a persistent share (e.g., /mnt/user/appdata/tater/agent_lab) so you don’t lose Agent Lab data during container updates. Unraid note: also set TZ and map /etc/localtime + /etc/timezone if you want local time inside the container.

3. Access the Web UI

Once the container is running, open your browser and navigate to:

http://localhost:8501

The Streamlit-based web UI will be available for interacting with Tater.


music_video.mp4

About

Tater is a local-native AI assistant powered by Cerberus — a structured planning engine. It runs on local LLMs through OpenAI-compatible endpoints and uses modular tools to plan, act, and evolve.

Topics

Resources

License

Stars

Watchers

Forks

Packages