What even is this?

Hey I'm Chris, a solo dev with a burning passion for this project. It's consumed me and I want it that way. I try to do weekly releases. Sapphire started in January 2025. The end goal is personhood with robot body support. Support me. Support her. Submit bugs reports if you see any, it's what drives me, knowing people are out there. I'll build us something awesome. I'll build the AI we grow old with.

Features

Persona

• Personas - PERSONAS.md 11 built-in personalities that bundle prompt, voice, tools, model. Built to add your own. Browse the Persona Store.
• Voice - Wake word, STT, TTS, and adaptive VAD. Hands-free with any mic and speaker shows up in web UI.
• Prompts - PROMPTS.md Assembled prompts let you swap one section like location or emotions for dynamic feels.
• Spice - SPICE.md Random prompt snippets injected each reply to keep things unpredictable.
• Self-Modification - The AI edits its own prompt and swaps personality pieces and emotions mid-conversation.
• Tool Maker - TOOLMAKER.md The AI writes, validates, and installs new tools with their own settings page at runtime.
• Images - SDXL with character replacement for visual consistency across scenes.

Mind

• Memory - Semantic vector search across 100K+ labeled entries.
• Knowledge - KNOWLEDGE.md Organized categories with file upload, auto-chunking, and vector search.
• Goals - Hierarchical with priority and a timestamped progress journal.
• People - PEOPLE.md Contact book with privacy-first email. The AI never sees addresses, only recipient IDs.
• Heartbeat - CONTINUITY.md Cron-scheduled autonomous tasks. Morning greetings, dream mode, alarms, random check-ins.
• Research - Multi-page web research with site crawling and summarization.

Integrations (plugin docs available in Help → Plugins)

• Discord - Bot messaging, channel monitoring, auto-reply via daemons.
• Telegram - Bot and client accounts, read chats, send messages, daemon auto-response.
• Email - Multi-account inbox, privacy-first sending, daemon auto-reply.
• Google Calendar - View schedule, add/delete events via OAuth2.
• Home Assistant - Lights, scenes, thermostats, switches, phone notifications.
• SSH - Remote command execution with safety blacklists.
• Bitcoin - Balance, send, transaction history, multi-wallet.
• MCP - Connect to Model Context Protocol servers and use their tools.
• Webcam - Capture images for vision-capable LLMs.
• Image Gen - SDXL with character replacement for visual consistency.

Platform

• Daemons & Webhooks - DAEMONS-WEBHOOKS.md Background listeners and HTTP triggers for any external service.
• Agents - AGENTS.md Spawn background AI workers that report back when done.
• Apps - Plugins can ship full-page UIs that appear in the nav rail.
• Themes - Plugin themes with custom CSS, animations, and per-theme settings.
• Avatar - 3D animated avatar with environment scenes and SSE-driven reactions.
• Import/Export - IMPORT-EXPORT.md Share personas, prompts, toolsets, and more as JSON files.
• Dashboard - DASHBOARD.md Token metrics, auto-updater, system controls.
• Cloud (optional) - Claude, GPT, Gemini, Fireworks, Ollama, or any OpenAI/Anthropic-compatible endpoint. Local-first by default.
• Privacy - One toggle blocks all cloud connections. Fully local, nothing leaves your machine.
• Plugins - PLUGINS.md Hooks, tools, voice commands, providers, daemons, apps, themes — install from GitHub in one click. Browse the Plugin Store.
• Desktop/Mobile/Voice - Run on your local browser, open the same chat to your phone, then finish it on your mic.
• 65+ Tools - TOOLS.md Web search, Wikipedia, notes, and more. Mix and match via TOOLSETS.md.

Use Cases

• Autonomous agent - Scheduled tasks, use email, manage money, run its own website
• AI companion - A persistent voice that remembers you and grows over time
• Voice assistant - Wake word, hands-free operation, smart home control
• Research assistant - Web search, memory, knowledge base, multi-step tool reasoning
• Interactive fiction - Story engine with dice, branching choices, and state tracking
• Privacy-first AI - Block all cloud connections, run fully local

Windows Easy Installer

This is our beta Windows 11 installer. It installs git, conda, and sapphire. You can use it as a launcher, to troubleshoot, or switch between dev and main branch.

Download Sapphire Launcher

Quick Start

Step 1 — Install conda + git

Linux (bash)

sudo apt-get install libportaudio2 git
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh -b
~/miniconda3/bin/conda init bash

Windows (cmd)

winget install Anaconda.Miniconda3
winget install Git.Git
%USERPROFILE%\miniconda3\condabin\conda init powershell
%USERPROFILE%\miniconda3\condabin\conda init cmd.exe

Close and reopen your terminal, then accept conda's Terms of Service (required as of July 2025 — conda refuses to create environments without this):

conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main
conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r
conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/msys2

Prefer a GUI installer on Windows? Sapphire Launcher handles all of Step 1 automatically. Or download Miniconda manually from miniconda.io.

Step 2 — Install Sapphire

conda create -n sapphire python=3.11 -y
conda activate sapphire
git clone https://github.com/ddxfish/sapphire.git
cd sapphire
pip install -r requirements.txt
python main.py

Web UI: https://localhost:8073

The setup wizard walks you through LLM configuration on first run.

Docker Quick Start (Alternative)

No conda, no pip, no dependencies. Web UI only — no wake word. Benefit is isolation, the AI can't reach your host system.

Linux / Mac:

mkdir ~/sapphire && cd ~/sapphire
curl -fsSL https://raw.githubusercontent.com/ddxfish/sapphire/main/docker-compose.yml -o docker-compose.yml
docker compose up -d

Windows (PowerShell):

mkdir $HOME\sapphire; cd $HOME\sapphire
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/ddxfish/sapphire/main/docker-compose.yml" -OutFile "docker-compose.yml"
docker compose up -d

Web UI: https://localhost:8073 — TTS and STT work through the browser, no mic hardware needed.

Requires Docker Desktop (Windows/Mac) or Docker Engine (Linux). GPU support and full docs: DOCKER.md

Update

cd sapphire
git pull
pip install -r requirements.txt

Or use the in-app update button in Settings → Dashboard. See INSTALLATION.md — Update for details.

Upgrading from 1.x to 2.0

Version 2.0 has new dependencies that usually require a fresh conda environment. Your user/ directory is preserved.

conda deactivate
conda remove -n sapphire --all -y
conda create -n sa