🧠
Multi-Model Intelligence
Connect to Gemini, ChatGPT, Claude, Grok, Kimi, Mistral, NVIDIA, GitHub
Copilot, Ollama, Apple Foundation Models, Prism Hosted, and custom OpenAI-compatible APIs.
Mix cloud, local, and on-device models without leaving Prism—including streaming chat on
Prism
Hosted when your account and plan include access.
Prism Hosted
GPT-5.5
Claude Opus 4.7
Gemini 3.1 Pro
Gemini 3 Flash
Gemini 3.1 Flash Lite
GPT-5.4 Mini
GPT-5.4
Claude 4.6 Sonnet
Qwen3.6 Flash
DeepSeek V4 Pro
GPT-5.4 Image 2
Kimi K2.6
GLM 5.1
Gemini
ChatGPT
Claude
Grok
Kimi
Mistral
GitHub Copilot
Llama 3
Apple Intelligence
NVIDIA
Custom APIs
✍️
System-Wide Writing
Transform any text field on your Mac into an AI workspace. Inline
autocomplete, keyboard-driven suggestion acceptance, and text injection workflows — all
powered
by macOS Accessibility APIs.
⚡
Quick AI Panel
A Spotlight-like floating bar you can summon from anywhere with Ctrl +
Space.
Ask a question, attach files, use slash commands, open the @ tool palette
for creators and MCP, browse custom web views, and dismiss just
as
fast. Always synced with the main window.
🔀
Model Comparison
Send the same prompt to multiple models simultaneously and compare
responses
side-by-side. Add unlimited slots, track generation speed, then synthesize the best answer
with
AI.
🎨
Image Generation
Generate 4K images with multiple art styles — Anime, Watercolor, Vector,
Illustration. Local generation via Ollama for complete privacy. Persistent gallery for all
your
creations.
📐
Advanced Math
Full LaTeX rendering with KaTeX for complex block equations and seamless
inline math. Automatic Greek letter conversion, fractions, roots, and boxed answers.
🌐
Browser Automation
An embedded Node.js server runs Playwright or Puppeteer so the AI can
navigate, scrape, and interact with the web agentically. Live WebSocket stream included.
🧩
Chrome Extension
Bring Prism's intelligence into Chrome. Enhance web pages, extract page
context, and enable seamless agentic browser control from the AI.
🔌
MCP Registry & @-Mention Tools
Configure Model Context Protocol servers from a curated registry—HTTP
and local transports, optional bearer tokens, and quick filtering. In chat, type
@ to open the tool palette: built-in @FileCreator,
@QuizCreator, and @FlashCardCreator run full creation
flows with streaming progress; MCP entries connect using your saved setup, discover remote
tools, and return structured context (including helpers like GitHub repository listing when
your prompt matches).
MCP catalog
@FileCreator
@QuizCreator
@FlashCardCreator
Tool discovery
❓
Quiz Me Mode
Enter any topic and get an AI-generated multiple-choice quiz with
configurable difficulty and length. Mix MCQ, true/false, and free-response questions with AI
grading and detailed explanations. When you start a run from chat with
@QuizCreator, the assistant message links straight to that session for
one-tap review.
🗂️
Folder Context Agent
Point Prism at a folder and it will scan files, extract text from PDFs,
build context windows, answer repo questions, and propose concrete file changes and terminal
commands.
📄
File Creator
Generate polished PDFs and editable files including Markdown, DOCX, HTML,
code, JSON, CSV, and more. Preview, revise, export, and keep a persistent document library.
Trigger full runs from chat with @FileCreator when you want the same flow
without leaving the thread.
🧠
Flash Cards
Create and study AI-generated decks with SM-2-style spaced repetition
(ease, interval, next review), deck management, provider and model labels, and a focused
study UI. @FlashCardCreator in chat saves the deck and links it on the
assistant message for quick reopen.
⌨️
Slash Commands
Type / for instant access to built-in commands like /summarize, /explain,
/translate, /fix, and /code. Create fully custom prompt templates with icons, plus action
commands like /new, /clear, and /quit.
🔍
Web Search & Reasoning
Enable real-time web search so the AI always has current information.
Adjust
thinking depth per provider, inspect reasoning output, and use multimodal prompts with
images,
PDFs, and pasted text attachments.
✨
AI Autocomplete
Get system-wide inline completion in text fields across macOS. Choose the
backend, model, debounce, completion length, custom instructions, memory, and app blacklist.
🪟
Web Overlay & Custom Web Views
Open Gemini, ChatGPT, Perplexity, Grok, Claude, or your own custom sites
in
a floating browser overlay with persistent sessions, private mode, and saved panel sizing.
🍎
Apple Shortcuts Integration
Route chat and image generation through your own Apple Shortcuts for
Siri,
private cloud, on-device, or ChatGPT-powered flows.
📎
Attachment-Aware Chat
Drag in images, PDFs, source files, CSV, JSON, and text files, paste
screenshots, use voice dictation where supported, and keep those attachments with the
conversation for later retries, edits, and follow-up prompts.
☁️
Chat History & iCloud
Chats live on disk by default. Optionally enable Sync Chat
History with iCloud across Macs signed into the same Apple ID. In Data &
Privacy you can also limit how many saved sessions Prism keeps—when
enabled,
oldest chats roll off automatically so storage stays predictable.
🔒
Private by Design
With Ollama or Apple Intelligence, prompts stay on your Mac. BYO API keys
talk directly to the providers you configure. Prism Hosted uses Prism’s
managed endpoints tied to your account when you pick that provider—no separate vendor API
console required for that path.