Skip to main content

​
What Are Assistants?

Assistants are specialized AI configurations designed for specific tasks, roles, or workflows. Instead of using a general-purpose AI for everything, assistants provide focused expertise with custom instructions, tools, and personality.
Think of assistants as expert colleagues - each one specialized in a different area, ready to help with their specific domain.

​
Why Use Assistants?

Specialized Expertise

Assistants are configured with domain-specific knowledge and instructions

Consistent Behavior

Get predictable, reliable responses tailored to your workflow

Time Savings

No need to repeat context or instructions in every conversation

Team Collaboration

Share assistants across your team for consistent results

Tool Optimization

Each assistant can have specific tools enabled for their task

Context Memory

Assistants remember their role and maintain appropriate context

​
Built-in vs Custom Assistants

Pre-configured by ZeroTwoZeroTwo includes ready-to-use assistants for common tasks:💻 Code Assistant:
  • Software development and debugging
  • Code review and optimization
  • Technical documentation
  • Architecture advice
✍️ Writing Assistant:
  • Content creation and editing
  • Copywriting and marketing
  • Blog posts and articles
  • Email and communication
📊 Data Analyst:
  • Data analysis and visualization
  • Statistical analysis
  • SQL and database queries
  • Report generation
🎨 Creative Assistant:
  • Brainstorming and ideation
  • Image generation guidance
  • Design feedback
  • Creative project planning
📚 Research Assistant:
  • Deep research and fact-finding
  • Literature review
  • Citation and references
  • Summary and synthesis
🏢 Business Assistant:
  • Business analysis
  • Strategy and planning
  • Market research
  • Presentation creation
Built-in assistants are available to all users and provide a great starting point.

​
How Assistants Work

Assistants combine several elements to provide specialized help:

​
Core Components

The assistant’s “personality” and roleSystem instructions define:
  • Who the assistant is
  • What they’re expert at
  • How they should respond
  • What they should prioritize
Example for Code Assistant:
You are an expert software engineer with 15+ years of experience.
You specialize in modern web development, particularly React, Node.js,
and TypeScript. When reviewing code, you focus on:

1. Code quality and maintainability
2. Performance optimization
3. Security best practices
4. Clear documentation

Always provide specific, actionable feedback with code examples.
Custom knowledge and contextUpload documents that the assistant references:
  • Company documentation
  • Style guides
  • Technical specifications
  • Process documents
  • Training materials
Examples:
  • Brand guidelines for writing assistant
  • API docs for coding assistant
  • Product catalog for sales assistant
  • Policy documents for compliance assistant
Which tools the assistant can useEnable specific tools for each assistant:Code Assistant:
  • ✅ Code Interpreter
  • ✅ GitHub integration
  • ✅ Web Search
  • ❌ Image Generation
Writing Assistant:
  • ✅ Web Search
  • ✅ Canvas
  • ❌ Code Interpreter
  • ❌ Image Generation
Creative Assistant:
  • ✅ Image Generation
  • ✅ Image Editing
  • ✅ Canvas
  • ✅ Web Search
How the assistant communicatesConfigure:
  • Response length (concise vs detailed)
  • Tone (formal, casual, friendly)
  • Structure (bullet points, paragraphs)
  • Technical level (beginner, intermediate, expert)
Example configurations:Customer Support:
  • Empathetic and patient tone
  • Clear, simple language
  • Step-by-step instructions
  • Always offer to help further
Technical Expert:
  • Precise technical language
  • Detailed explanations
  • Code examples included
  • Assumes technical knowledge
Which AI model powers the assistantChoose the best model for the task:
  • GPT-4o: Fast, versatile, good for most tasks
  • Claude 3.5 Sonnet: Best for coding and long-form writing
  • o1: Deep reasoning for complex problems
  • Gemini 2.0 Flash: Quick responses with vision
Different tasks benefit from different models.

​
Using Assistants

​
Selecting an Assistant

1

Open assistant selector

Click the Assistant dropdown in the chat headerKeyboard shortcut: Cmd/Ctrl + K
2

Browse assistants

See available assistants:
  • ⭐ Starred assistants (top)
  • 📁 Your custom assistants
  • 🏢 Team assistants
  • 🌟 Built-in assistants
3

Select assistant

Click the assistant you want to useThe assistant’s name and icon appear in the header
4

Start conversing

Your messages now go to that specific assistant
The assistant uses its configured instructions, knowledge, and tools.

​
Switching Assistants

You can switch assistants mid-conversation:
1

Click assistant name

Click the current assistant name in the header
2

Choose new assistant

Select a different assistant from the list
3

Continue conversation

The new assistant sees the conversation history and continues from there
Context is maintained when switching assistants within the same conversation.

​
Default Assistant

Set a default assistant for new conversations: Settings > Assistants > Default Assistant Choose which assistant to use automatically when starting new chats.

​
Assistant Features

​
Memory & Context

Assistants maintain context across conversations: Session Memory:
  • Remembers current conversation
  • Tracks discussed topics
  • Maintains user preferences
  • Recalls previous decisions
Long-Term Memory (Pro/Enterprise):
  • Remembers past conversations
  • Learns from interactions
  • Maintains project context
  • Adapts to your style
Assistants with memory provide more personalized and context-aware help over time.

​
Conversation Starters

Pre-defined prompts to help users get started: Example for Code Assistant:
  • “Review this code for improvements”
  • “Help me debug this error”
  • “Explain this code snippet”
  • “Suggest optimizations”
Example for Writing Assistant:
  • “Help me write a blog post about…”
  • “Improve this draft”
  • “Write an email to…”
  • “Create an outline for…”

​
Assistant Actions

Quick actions assistants can perform:

Code Review

Automatically analyze uploaded code with specific criteria

Document Summary

Summarize uploaded documents in assistant’s style

Template Fill

Use assistant to fill predefined templates

Automated Responses

Generate responses in assistant’s voice

​
Assistant Categories

Assistants organized by use case:
Software engineering assistants
  • Frontend Developer: React, Vue, Angular expertise
  • Backend Developer: API, database, server logic
  • DevOps Engineer: CI/CD, infrastructure, deployment
  • Mobile Developer: iOS, Android, React Native
  • Full Stack: End-to-end development
  • QA Engineer: Testing, quality assurance
  • Security Expert: Code security, vulnerability review
  • Code Reviewer: PR reviews, code quality

​
Assistant Marketplace

Browse and install community-created assistants:

​
Finding Assistants

1

Open marketplace

Settings > Assistants > Marketplace
2

Browse or search

  • Popular assistants
  • Recently added
  • By category
  • Search by keyword
3

View details

Click assistant to see:
  • Description and capabilities
  • Creator and reviews
  • Required tools
  • Example conversations
4

Install

Click Install to add to your workspace

​
Publishing Your Assistants

Share your custom assistants with the community:
1

Create assistant

Build and test your custom assistant
2

Click Publish

Assistant settings > Publish to Marketplace
3

Add details

  • Name and description
  • Category
  • Example use cases
  • Screenshots
4

Submit

Submit for review (typically approved within 24 hours)

​
Team Assistants

Share assistants across your organization:

​
Sharing with Team

1

Open assistant settings

Select assistant > Settings icon
2

Click Share

Click Share with Team
3

Set permissions

  • Can use: Team members can use assistant
  • Can edit: Team members can modify assistant
  • Can duplicate: Team members can copy and customize
4

Share

Assistant appears in team’s assistant library

​
Organization Assistant Library

Enterprise feature: Central library of approved assistants Benefits:
  • Consistent assistants across organization
  • Quality control and approval workflow
  • Usage analytics
  • Version management
  • Compliance assurance
Access: Organization Settings > Assistants

​
Best Practices

Learn before building:
  1. Try built-in assistants first
  2. Understand what makes them effective
  3. Identify what you’d change
  4. Use as templates for custom assistants
Most needs are covered by built-in or lightly customized assistants.
Clear, specific instructions work best:✅ Good:
You are a React code reviewer. Focus on:
- Component structure and reusability
- Proper hook usage
- Performance optimization
- Accessibility compliance
❌ Too broad:
You help with coding and can review any code in any language 
and also help with writing and data analysis.
Validate before sharing:
  1. Test with various inputs
  2. Check edge cases
  3. Verify tool usage
  4. Confirm tone and style
  5. Get feedback from others
  6. Iterate based on results
Descriptive names help users:✅ Clear names:
  • “React Code Reviewer”
  • “SEO Blog Post Writer”
  • “SQL Query Assistant”
  • “Customer Support Agent”
❌ Unclear names:
  • “Helper”
  • “My Assistant”
  • “AI 1”
  • “Test”

​
Next Steps

Create Assistant

Build your first custom assistant

Configure Behavior

Fine-tune assistant personality and responses

Manage Assistants

Organize and maintain your assistants

View Examples

Get inspired by example configurations
Assistants supercharge your productivity by providing specialized expertise exactly when you need it!