Skip to content

USER-GUIDE.md

Latest commit

 

History

History
297 lines (187 loc) · 11.1 KB

USER-GUIDE.md

File metadata and controls

297 lines (187 loc) · 11.1 KB

MultAI — User Guide

Version: 0.3.0 | Date: 2026-03-18

What is MultAI?

MultAI is a research assistant that sends your questions to seven different AI services at the same time — Claude, ChatGPT, Copilot, Perplexity, Grok, DeepSeek, and Gemini — then collects all the answers and combines them into a single, polished report.

It runs as a plugin inside Claude Code, the command-line tool from Anthropic. You simply type your question in plain English, and MultAI takes care of everything: opening browsers, submitting prompts, waiting for answers, and synthesizing the results. No coding required.

Why use seven AIs instead of one? Every AI has blind spots. By comparing answers from multiple sources, you get a more complete, more balanced picture — with less risk of missing something important.

What Can MultAI Do?

Ask Any Question to 7 AIs at Once

The simplest way to use MultAI: just ask a question. MultAI sends it to all seven AI platforms simultaneously and gives you a combined answer.

Example: "What are the key trends in renewable energy for 2026?"

What you get: Answers from all 7 AIs, synthesized into a single report highlighting consensus, disagreements, and unique insights from each source.

Research a Product in Depth

Want to understand what a software product can really do? The Solution Researcher runs a deep competitive intelligence analysis across all seven AIs and produces a structured capability report.

Example: "Research Notion — I want to understand its project management features"

What you get: A Consolidated Intelligence Report covering capabilities, limitations, pricing, and how it compares to alternatives — backed by evidence from seven independent AI assessments.

Map an Entire Market

Need to understand a whole product category? The Landscape Researcher creates a comprehensive market landscape report with interactive charts.

Example: "Give me a market landscape of project management tools"

What you get: A 9-section market landscape report with leader/challenger classifications, trend analysis, and interactive comparison charts — all auto-generated from seven AI perspectives.

Compare Products Side by Side

Once you have research reports on multiple products, the Comparator creates and maintains a feature-by-feature spreadsheet comparing them all.

Example: "Add Asana to the comparison matrix"

What you get: An updated Excel spreadsheet with Asana's capabilities ticked against every feature in the matrix, complete with weighted scoring.

Combine AI Answers into One Report

The Consolidator merges raw AI answers into a single structured report. This usually runs automatically as part of the other workflows, but you can also invoke it directly.

Before You Start

What You Need

Requirement Details
Operating system macOS (recommended), Linux, or Windows
Python Version 3.11 or higher — download here
Google Chrome Latest version — download here
Claude Code Anthropic's CLI tool — install instructions

Log In to AI Platforms

MultAI uses your existing Chrome browser sessions to interact with AI platforms. Before your first run, open Chrome and sign in to each platform:

Platform URL Account
Claude.ai claude.ai Free Anthropic account
ChatGPT chat.openai.com Free OpenAI account
Microsoft Copilot copilot.microsoft.com Free Microsoft account
Perplexity perplexity.ai Free account
Grok grok.com Free X/Twitter account
DeepSeek chat.deepseek.com Free account
Google Gemini gemini.google.com Free Google account

Tip: You don't need to sign in to all seven. MultAI will use whichever platforms you're logged into and skip the rest.

Installation

Option A: Plugin Install (Recommended)

This is the easiest way to get started. One command does everything.

Step 1. Open your terminal (Terminal on macOS, or your preferred terminal app).

Step 2. Run:

claude plugin install alo-exp/multai

Step 3. That's it. Dependencies install automatically the first time you use MultAI.

Option B: Manual Install

If you prefer to manage the installation yourself, or if the plugin install didn't work:

Step 1. Open your terminal.

Step 2. Clone the repository:

git clone https://github.com/alo-exp/multai.git

Step 3. Enter the project directory:

cd multai

Step 4. Run the setup script:

bash setup.sh

Step 5. Wait for setup to complete (usually takes 1-2 minutes). You'll see progress messages as it installs the required packages and downloads the browser automation tools.

When you see Setup complete. You're ready to use MultAI. — you're done.

Optional: Enable Smart Recovery

MultAI includes a "smart recovery" feature that uses a backup AI to help when something goes wrong during a research run (for example, if a platform's layout has changed). This is optional but recommended.

To enable it, you need a free API key from one of these providers:

Option 1: Free Google Gemini API Key (Recommended)

  1. Go to aistudio.google.com/apikey
  2. Click Create API Key and copy the key
  3. Create a file called .env in the MultAI project folder and add: 
    GOOGLE_API_KEY=your_key_here

Option 2: Anthropic API Key

  1. Go to console.anthropic.com
  2. Create an API key and copy it
  3. Add to your .env file: 
    ANTHROPIC_API_KEY=your_key_here

Then run: bash setup.sh --with-fallback

Your First Research Run

Here's a step-by-step walkthrough of your first MultAI run:

Step 1. Open Claude Code in the MultAI project directory:

cd multai
claude

Step 2. Type your research question naturally. For example:

What are the pros and cons of using Kubernetes for small teams?

Step 3. MultAI automatically detects that this is a multi-AI research question and submits it to all seven platforms. You'll see progress messages as each platform responds:

[Claude.ai] Navigating to https://claude.ai/new
[ChatGPT] Navigating to https://chat.openai.com
[Perplexity] Navigating to https://www.perplexity.ai
...

Step 4. After all platforms respond (usually 2-5 minutes for a regular question), MultAI synthesizes the answers into a single combined report. The report opens automatically in the interactive viewer.

Tip: If you want deeper, more thorough research, include the word "deep" in your request: "Do a deep research on Kubernetes for small teams"

Viewing Your Reports

The Report Viewer

MultAI includes a beautiful interactive report viewer that displays your research results. It opens automatically when a report is ready.

What you'll see:

  • Top navigation bar — Click vendor/section names to jump directly to that section
  • Left sidebar — Full table of contents; click any heading to navigate
  • Main content area — Your research report with formatted text, tables, and interactive charts
  • Export buttons (bottom right):
    • 📋 Copy — Copies the report for pasting into Google Docs or other editors
    • 📄 PDF — Exports the report as a PDF file

Where Reports Are Saved

All reports are saved in the reports/ folder inside the MultAI project directory. Each research run creates its own subfolder:

reports/
├── my-kubernetes-research/
│   ├── Claude.ai-raw-response.md       (individual AI answers)
│   ├── ChatGPT-raw-response.md
│   ├── ...
│   ├── status.json                      (run metadata)
│   └── my-kubernetes-research - Raw AI Responses.md  (combined archive)

DEEP vs. REGULAR Mode

MultAI has two research modes:

REGULAR Mode DEEP Mode
Speed 2-5 minutes 10-30 minutes
Depth Quick, focused answers Thorough research with web browsing
When to use Quick questions, fact-checking Market research, competitive intelligence
How to request Just ask your question Include "deep" in your request

Example (REGULAR): "What is Terraform?"

Example (DEEP): "Do a deep research on Terraform's enterprise features"

Tips and Tricks

Run fewer platforms for speed. If you only need a quick answer, you can ask MultAI to use just a few platforms:

"Use only Claude and ChatGPT to answer: What is Infrastructure as Code?"

Check your budget before a big run. Each AI platform has usage limits. Ask MultAI to show your remaining budget:

"Show me the rate limit budget"

Re-run a single failed platform. If one platform didn't respond, you can re-run just that one:

"Run the same question on just Perplexity"

DEEP mode on ChatGPT and Claude takes time. Deep Research on these platforms can take 10-20 minutes. This is normal — they're doing thorough web research behind the scenes.

Reports are cumulative. Each run creates a new folder. Your old reports are never overwritten or deleted.

Troubleshooting

Chrome won't connect

What happened: MultAI couldn't connect to your Chrome browser.

Fix: Close all Chrome windows completely, wait a few seconds, then try again. If that doesn't work, the research run will automatically launch a new Chrome instance.

Rate limited

What happened: You've used up your free quota on one or more AI platforms.

Fix: Wait a while (usually a few hours), or check your budget with: "Show me the rate limit budget." MultAI automatically skips platforms that are over their limit.

A platform returned no results

What happened: One of the seven AIs didn't respond or had an error.

Fix: This is normal — the other platforms' answers are still used. You can re-run just the failed platform if you want its input.

Report viewer shows a blank page

What happened: The report viewer loaded before the file was ready, or the server isn't running.

Fix: Refresh the page. If that doesn't work, ask MultAI to reopen the report.

Further Reading

Resource Description
Help Center Live documentation — installation, concepts, CLI reference, and troubleshooting
CONTRIBUTOR-GUIDE.md Technical reference for contributors — CLI details, platform internals, adding new platforms, tests, CI/CD
GitHub Repository Source code, issues, and releases
Claude Code Documentation How to use Claude Code (the tool MultAI runs inside)

MultAI is open source. Built with Python, Playwright, and Claude Code.