Setup Guides

Step-by-step walkthroughs for every integration. Follow along — no prior experience required.

Gmail Setup

Time: ~5 minutes. You need a Google account and a web browser.

1

Go to Google Cloud Console

Open console.cloud.google.com. Sign in with the Google account you want to use with PRISM.

2

Create a new project

Click the project dropdown at the top → New Project. Name it "PRISM" or anything you like. Click Create.

3

Enable the Gmail API

Navigate to APIs & Services → Library. Search for "Gmail API". Click it, then click Enable.

4

Set up the OAuth consent screen

Go to APIs & Services → OAuth consent screen.

  • Choose External user type
  • App name: "PRISM" (or anything)
  • User support email: your email
  • Developer contact: your email
  • Click Save and Continue through all steps
  • On the Test users screen, add your Gmail address
5

Create OAuth credentials

Go to APIs & Services → Credentials → Create Credentials → OAuth client ID.

  • Application type: Desktop app
  • Name: "PRISM"
  • Click Create
  • Click Download JSON on the confirmation dialog
6

Connect PRISM

# Move the downloaded JSON to PRISM's config directory
mkdir -p ~/.prism
mv ~/Downloads/client_secret_*.json ~/.prism/credentials.json
# Authenticate (opens your browser)
prism auth

Your browser opens to Google's sign-in page. Approve access. You'll see "Authentication successful" in the terminal. Done.

Note: While your app is in "Testing" mode, tokens expire after 7 days of inactivity. Just run prism auth again. To make tokens permanent, publish your app in Google Cloud Console (it's free, just requires a brief review).

📩

Outlook Setup

Outlook support uses Microsoft Graph API. Requires a Microsoft account.

Outlook setup documentation is being finalized. The provider is built and functional — it uses MSAL (Microsoft Authentication Library) for OAuth and connects to Microsoft Graph for all email operations.

For early access, see the setup guide on GitHub.

C

Claude API Key (Anthropic)

What is BYOK? "Bring Your Own Key" means you provide your own AI API key. PRISM never stores your key on any server — it's encrypted locally on your machine with AES-256. You pay Anthropic directly at their rates. PRISM has zero access to your AI usage data.

1

Go to the Anthropic Console

Open console.anthropic.com. Create an account or sign in.

2

Add billing

Go to Settings → Billing. Add a payment method. You can set a monthly spend limit to control costs.

3

Create an API key

Go to Settings → API Keys. Click Create Key. Name it "PRISM" (optional). Copy the key — it starts with sk-ant-.

Important: Copy the key now. Anthropic only shows it once. If you lose it, create a new one.
4

Add it to PRISM Dashboard

Open the PRISM dashboard → Settings → AI Chat. Select "Claude" as your provider. Paste your key. Click Save.

Your key is encrypted with AES-256-GCM and stored only on your machine.

Recommended models

Claude Haiku 4.5 — Fast, cheap, great for quick email tasks ~$0.25/1M input tokens
Claude Sonnet 4.6 — Balanced speed and intelligence ~$3/1M input tokens
Claude Opus 4.6 — Most capable, best for complex email workflows ~$15/1M input tokens
O

OpenAI API Key

1

Go to the OpenAI Platform

Open platform.openai.com. Sign in with your OpenAI account.

2

Add billing

Go to Settings → Billing. Add credits (minimum $5). You can set a monthly budget limit.

3

Create an API key

Go to API Keys (in the left sidebar). Click Create new secret key. Name it "PRISM". Copy the key — it starts with sk-.

4

Add it to PRISM Dashboard

Open the PRISM dashboard → Settings → AI Chat. Select "OpenAI" as your provider. Paste your key. Click Save.

Recommended models

GPT-4o mini — Fast and cheap, good for simple email tasks ~$0.15/1M input tokens
GPT-4o — Best balance of speed, cost, and capability ~$2.50/1M input tokens

Ollama (Local AI)

100% local, 100% free. Ollama runs AI models on your own hardware. No API key needed. No data leaves your machine. Perfect for maximum privacy.

1

Install Ollama

Download from ollama.com. Available for macOS, Linux, and Windows.

2

Pull a model

# Good balance of speed and quality (~4GB)
ollama pull llama3.2
# Or for better tool calling (~8GB)
ollama pull qwen2.5:14b
3

Start Ollama

ollama serve

Ollama runs on http://localhost:11434 by default.

4

Configure in PRISM Dashboard

Open the PRISM dashboard → Settings → AI Chat. Select "Ollama" as your provider. No API key needed — just select your model from the dropdown.

Hardware note: Local models need decent hardware. Minimum 8GB RAM for small models, 16GB+ recommended for larger ones. A GPU significantly improves speed but isn't required.

💰

AI Cost Guide

PRISM's AI chat is BYOK — you pay your provider directly. Here's what to expect for typical email usage.

Usage Level Description Claude Haiku GPT-4o mini Ollama
Light 10-20 queries/day ~$1-3/mo ~$1-2/mo $0 (free)
Moderate 50-100 queries/day ~$5-10/mo ~$3-7/mo $0 (free)
Heavy 200+ queries/day ~$10-20/mo ~$8-15/mo $0 (free)

Prompt caching reduces costs by ~90% for repeated context. PRISM caches your email context between messages so you're not re-sending your entire inbox every turn.

Tip: Start with Claude Haiku or GPT-4o mini. They're fast, cheap, and handle most email tasks perfectly. Upgrade to larger models only when you need complex multi-step workflows.

Budget controls: Both Anthropic and OpenAI let you set monthly spend limits in their dashboards. Set one before you start.