🧭 Instructions for Radiologists

🔹 Overview

RadCloudAI is a secure, AI-powered platform built for radiologists who want to streamline structured reporting, automate QA, and stay up to date with literature — all while retaining full control over their data and model selection.

It supports both image-based and text-based inputs and delivers instant outputs with no installation required.

🔹 What You’ll Need

To start using RadCloudAI, you’ll need:

✅ A RadCloudAI account (free to create)
✅ A valid API key or service account from a supported LLM provider (see below)
✅ A desktop browser — no special software required

🔹 Supported Inputs

You can generate reports using:

Free-text reports✅ Yes

Screenshots (JPG/PNG)✅ Yes

Dictation paste-ins✅ Yes

Combo (text + image)✅ Yes

You can paste reports from dictation systems (e.g., Fluency), or drag-and-drop PACS screenshots into the interface.

🔹 Core Features

FeatureDescription

Full Report GeneratorBuilds a full radiology report (Findings, Impression, Recommendations)

Impression GeneratorGenerates just the final impression from dictated or captured findings

Literature Review (Live)PubMed and Radiopaedia links extracted per caseCME IntegrationSee relevant CME links (STATdx, UTD, Radiopaedia) on every output

Automated QARule-based and AI-powered QA alerts per caseImage SupportSubmit screenshots for interpretation using vision models (e.g., Gemini)

Text SupportPaste your dictated reports directlyGroup-Based ConfigurationEach group uses its own LLMs, routing logic, and credentials

Peer Review (Coming Soon)Built-in anonymous peer scoring and commentary (launching shortly)

🔐 Bring Your Own LLM Key

To use RadCloudAI, you must connect an account with an AI provider.

Supported LLM providers include:

✅ OpenAI

  • GPT-4o, GPT-4-turbo, GPT-3.5

  • Create account at platform.openai.com

  • Paste API key into Settings → Credentials

✅ Google Vertex AI (Gemini)

  • Gemini 2.5 Pro for text + image workflows

  • Use service account JSON key (requires Google Cloud project)

We also support Anthropic, Mistral, Cohere, Perplexity, and others — you’ll need to bring your own credentials.

⚙️ Don’t Want to Manage the API Setup?

We offer optional paid LLM connection services, including:

  • Key provisioning

  • Usage monitoring

  • Support for credential rotation

  • Group onboarding

📧 Contact support@radcloudai.org
💳 Pricing is usage-based and scaled per user group

🔍 Tips for Clinical Use

  • Use image-only workflows for screenshots of limited dictations

  • Use text-only mode if you already dictated the case

  • Use combo mode (text + image) for longer, multi-phase exams

  • View QA summary on every case — color-coded by severity

  • Copy/paste literature and CME links into your reporting system or teaching file

🧠 Who This Is For

This platform is designed for:

  • Radiologists comfortable with API configuration

  • Group leaders managing LLM-based workflows

  • Academic users doing QA or guideline adherence tracking

  • IT-savvy clinicians seeking independence from vendor lock-in

🛠️ Support

If you need help with:

  • LLM setup (OpenAI, Gemini)

  • Group creation

  • Credential encryption

  • CME link behavior

  • QA feedback tuning

📧 Reach us anytime at support@radcloudai.org

✅ Ready to Begin?

Once your LLM key is entered, head to:

👉 Create Report
👉 Settings → Model Configuration to personalize your routing

No patient data is stored, and all keys are encrypted per group.