ChatNGO Labs Onboarding

SimTheory Platform Guide

 

1. Platform Basics

SimTheory provides access to multiple AI models with enhanced functionality beyond standard chat interfaces.

Model Selection

  • Frontier models: GPT-4, Claude Sonnet, Gemini 2.5 Pro (count towards usage)
  • Efficient alternatives: Kimi K2, DeepSeek, Claude 3.5 Haiku (no usage cost)
  • Switch models based on task complexity

Usage Management

  • Monitor tokens via Profile → Settings → Usage
  • Three types: Input tokens, Output tokens, Premium tokens
  • Monthly caps apply per user

2. AI Assistants

Create specialized AI agents for different organizational tasks and workflows.

Assistant Types

  • SimTheory: Default assistant with memory enabled
  • Blank: Memory disabled for temporary conversations
  • Custom: Task-specific assistants with specialized instructions

Configuration Options

  • Upload context files (strategic plans, style guides)
  • Set custom instructions and default models
  • Enable specific MCP tools per assistant
  • Use pre-built templates for quick setup

3. MCP Integration

MCPs are universal connectors that link your AI to external tools and services.

Key MCP Categories

  • Research: Academic databases, Google Search, Perplexity, web crawling
  • Productivity: Google Workspace, code interpretation, data analysis
  • Creative: Podcast generation, image creation and editing
  • Data: Government datasets, mapping, geographic analysis

MCP Management

  • Install via “Add Connection”
  • Enable/disable per assistant
  • Custom naming for organization
  • Workspace sharing capabilities

4. Advanced Workflows

Leverage multiple AI agents and tools simultaneously for complex organizational tasks.

Multi-Tab Operations

  • Run different AI agents simultaneously
  • Switch between projects seamlessly
  • Organize conversations in folders
  • Pin important chats for quick access

Dynamic Switching

  • Change models mid-conversation
  • Switch assistants during workflow
  • Optimize for different task requirements

5. Best Practices

Maximize platform efficiency while managing costs and complexity.

Strategic Usage

  • Use frontier models for complex tool calling
  • Switch to efficient models for basic tasks
  • Monitor token consumption regularly
  • Restrict MCP access to relevant tools only

Workflow Optimization

  • Create specialized assistants for recurring tasks
  • Run research agents in background tabs
  • Chain different assistants for complex workflows
  • Balance memory-enabled and private sessions