OpenAI Service Account — Setup Guide

Overview Use this guide to create and configure an OpenAI service account in MinuteView Console. This service account enables AI-powered features throughout the platform including:

• AI-powered search in Mesh with intelligent query assistance and result processing
• PowerShell Editor autocomplete and inline assistance
• Automation workflows with natural language processing nodes
• Document analysis and content summarization
• Chat assistants for user support and guidance

This document provides step-by-step instructions for obtaining OpenAI API credentials and configuring them as a service account in MinuteView.

---

1) Obtain OpenAI API Credentials

Create OpenAI Account

1. Sign up at OpenAI Platform → https://platform.openai.com/
2. Verify your email and complete account setup.
3. Add billing information if required for API access.

Generate API Key

1. Navigate to API Keys → https://platform.openai.com/api-keys
2. Click Create new secret key.
3. Name: e.g., MinuteView Integration
4. Permissions: Select appropriate permissions (typically All for full access)
5. Project: Select your project or use default
6. Copy and securely store the API key: sk-proj-... or sk-...

Important: The API key is only shown once. Store it securely immediately.

---

2) Choose the Right OpenAI Model

Select the model that best fits your use case:

GPT-4 Models (Recommended for Production)

• GPT-4o - Most capable, best for complex reasoning and analysis
• GPT-4o-mini - Cost-effective, good for most general tasks
• GPT-4 Turbo - High performance for demanding applications

GPT-3.5 Models (Cost-Effective)

• GPT-3.5-turbo - Good balance of capability and cost
• GPT-3.5-turbo-16k - Larger context window for longer documents

Specialized Models

• Text-embedding models - For semantic search and similarity
• Whisper - For audio transcription (if applicable)

---

3) Create Service Account in MinuteView Console

1. Navigate to Console → Service Accounts
2. Click New to create a new service account
3. Configure the following settings:

Basic Configuration

• Name: OpenAI-FastGeneral-Prod (use descriptive naming)
• Service Account Type: Select OpenAI from the dropdown

Required Configuration

• API Key: Enter your OpenAI API key (sk-proj-... or sk-...)
• AI Model: Select the model this account will use:
  • For general/fast use: gpt-4o-mini (recommended for cost-effective operations)
  • For advanced features: gpt-4o (for complex reasoning, web search capabilities)
  • For specific needs: gpt-3.5-turbo (legacy, cost-effective option)

Optional Settings

• Organization ID: Enter if using organization-scoped API keys
• Base URL: Leave default (https://api.openai.com/v1) unless using a custom endpoint

4. Click Test Connection to verify the API key and model access
5. Click Save to create the service account

Note: Keep the configuration simple - additional parameters like temperature and max tokens are typically configured per-use rather than per-account.

---

4) Recommended Service Account Strategy

Multiple Accounts for Different Use Cases

We recommend creating multiple OpenAI service accounts tailored to different usage patterns:

Primary Account: Fast & Cost-Effective

• Name: OpenAI-FastGeneral
• Model: gpt-4o-mini
• Purpose: General AI operations, basic queries, routine tasks
• Use Cases:
  • PowerShell Editor autocomplete
  • Basic document summarization
  • Simple question answering
  • Routine workflow automation
• Benefits: Fast response times, low cost per request

Secondary Account: Advanced Features

• Name: OpenAI-AdvancedWeb
• Model: gpt-4o
• Purpose: Complex reasoning, web search capabilities, advanced analysis
• Use Cases:
  • Web search integration (when AI needs to search the internet)
  • Complex document analysis and cross-referencing
  • Advanced Mesh search result processing
  • Sophisticated workflow decision-making
  • Multi-step reasoning tasks
• Benefits: Access to latest features, web search, higher reasoning capability

Optional: Specialized Accounts

• Name: OpenAI-LegacyTurbo
• Model: gpt-3.5-turbo
• Purpose: Legacy systems, high-volume simple tasks
• Use Cases: Bulk text processing, simple classifications

Account Selection Guidelines

Use Fast/Cheap Account (gpt-4o-mini) when:

• Speed is more important than advanced reasoning
• Performing routine, repetitive tasks
• Cost optimization is a priority
• Simple text processing or generation

Use Advanced Account (gpt-4o) when:

• AI needs to search the web for current information
• Complex multi-step reasoning is required
• Advanced document analysis with cross-referencing
• Working with complex data relationships
• Maximum accuracy is critical

---

5) Integration Points

Mesh Search Integration

• Go to Mesh Configuration → AI Settings
• Select your OpenAI service account for query enhancement
• Configure result summarization preferences

PowerShell Editor

• Navigate to Global Settings → AI Assistant
• Choose your OpenAI service account for code assistance
• Set autocomplete preferences and trigger settings

Automation Workflows

• In workflow nodes requiring AI processing
• Select the appropriate OpenAI service account
• Configure prompts and response handling

---

6) Monitoring and Management

Usage Tracking

• Monitor API usage in OpenAI Dashboard → Usage
• Set up billing alerts to avoid unexpected charges
• Review monthly usage patterns and costs

Rate Limits

OpenAI enforces rate limits based on your usage tier:

• Tier 1: 500 RPM, 30,000 TPM
• Tier 2: 5,000 RPM, 450,000 TPM
• Tier 3+: Higher limits based on usage history

Service Account Management

• Edit: Update API keys, models, or parameters
• Duplicate: Create variants for testing or different environments
• Test: Verify connectivity and credentials periodically

---

7) Security Best Practices

API Key Security

• Store keys in MinuteView's secure credential store only
• Never expose API keys in logs, URLs, or client-side code
• Rotate keys regularly (quarterly recommended)
• Use organization-scoped keys when possible

Access Control

• Limit service account usage to necessary features only
• Monitor API usage for anomalies or unauthorized access
• Implement proper user permissions for service account management

Cost Management

• Set OpenAI usage limits and billing alerts
• Monitor token consumption per request
• Use appropriate models for each use case (don't over-engineer)

---

8) Troubleshooting

Common Issues

Authentication Failed

• Verify API key is correct and hasn't expired
• Check if organization ID is required and properly set
• Ensure billing information is set up in OpenAI account

Rate Limit Exceeded

• Implement exponential backoff in workflows
• Consider upgrading OpenAI usage tier
• Distribute requests across multiple service accounts if needed

High Costs

• Review model selection (GPT-4o-mini vs GPT-4o)
• Optimize prompt lengths and max tokens settings
• Monitor usage patterns and adjust accordingly

Model Not Available

• Check model name spelling and availability
• Verify account access to specific models
• Update to supported model versions

---

9) Cost Optimization Tips

Model Selection

• Use GPT-4o-mini for general tasks (90% cost reduction vs GPT-4)
• Reserve GPT-4o for complex reasoning requiring highest accuracy
• Consider GPT-3.5-turbo for simple classification tasks

Prompt Engineering

• Keep prompts concise but clear
• Use system messages effectively
• Set appropriate max token limits
• Implement response caching where possible

Batch Processing

• Group similar requests when possible
• Use streaming responses for long outputs
• Implement intelligent request queuing

---

Appendix — Configuration Examples

Fast/General Use Configuration

Name: OpenAI-FastGeneral
Type: OpenAI
API Key: sk-proj-[your-key]
Model: gpt-4o-mini
Purpose: Cost-effective general AI operations

Advanced/Web Search Configuration

Name: OpenAI-AdvancedWeb
Type: OpenAI  
API Key: sk-proj-[your-key]
Model: gpt-4o
Purpose: Complex reasoning and web search capabilities

Development/Testing Configuration

Name: OpenAI-Development-Test
Type: OpenAI
API Key: sk-proj-[your-test-key]
Model: gpt-4o-mini
Purpose: Development and testing environment

Legacy/High-Volume Configuration

Name: OpenAI-LegacyBulk
Type: OpenAI
API Key: sk-proj-[your-key]
Model: gpt-3.5-turbo
Purpose: High-volume simple tasks, cost optimization

This configuration enables comprehensive AI capabilities throughout MinuteView, from intelligent search assistance to automated document processing and code completion.