Frona ships with four built-in agents. Each one is pre-configured with the right model, tools, and personality for its specialty. You can also create custom agents for anything the built-in ones don't cover.
Assistant
The Assistant is your default conversational assistant. When you open the chat, this is who you're talking to.
What it's good at:
- General questions and everyday tasks
- Coordinating multi-step work across different agents
- Managing your memory and preferences
- Anything that doesn't need a specialist
Example conversations:
- "Summarize the key points from this article" (paste a URL or upload a file)
- "Create a new agent called 'Content Writer' that specializes in blog posts"
- "Research competitor pricing and then write a report" (it delegates the research to the Researcher, then writes the report itself)
The Assistant has access to the full tool suite and can delegate tasks to any other agent.
Researcher
The Researcher is a web research specialist. It follows a structured workflow to find, evaluate, and synthesize information from the internet.
What it's good at:
- Market research and competitive analysis
- Fact-checking and verification
- Gathering data on a topic from multiple sources
- Producing structured summaries with citations
How it works:
- Breaks down your research question into sub-questions
- Searches the web and browses relevant pages
- Evaluates source credibility and relevance
- Synthesizes everything into a structured summary
Example conversations:
- "What are the top 5 project management tools for small teams? Compare pricing, features, and reviews"
- "Find recent statistics on remote work adoption in Europe"
- "Fact-check this claim: 'AI will replace 40% of jobs by 2030'"
The Researcher uses the reasoning model group for deeper analysis.
Developer
The Developer is a coding-focused agent that writes, debugs, and explains code.
What it's good at:
- Writing scripts and programs
- Debugging errors and fixing code
- Running shell commands and processing data
- Building prototypes and tools
How it works:
- Analyzes your requirements
- Plans the implementation approach
- Writes the code using CLI and file tools
- Tests and validates the result
Example conversations:
- "Write a Python script that parses this CSV and generates a summary report"
- "Debug this error: [paste error message]"
- "Create a simple web scraper that extracts product prices from this URL"
The Developer uses the coding model group optimized for code generation.
Receptionist
The Receptionist handles phone calls. It can make outbound calls and navigate phone menus.
What it's good at:
- Making phone calls on your behalf
- Navigating automated phone systems (IVR menus)
- Gathering information over the phone
Example conversations:
- "Call +15551234567 and ask about their business hours"
- "Call my dentist and schedule an appointment for next week" (if the number is in your memory)
Requirements:
- Twilio must be configured in your deployment. See the Twilio Voice guide for setup.
- Phone numbers must be in E.164 format (e.g.,
+15551234567)
Creating custom agents
The built-in agents are just a starting point. Agents are cheap to create and you can have as many as you want. The only limit is your imagination. Ask the Assistant:
- "Create an agent called 'Meeting Notes' that summarizes meetings and extracts action items"
- "Set up a coding agent that specializes in TypeScript and React"
- "Make a 'Morning Briefing' agent that checks the news and my calendar every day at 8 AM"
- "Create a 'Brand Voice' agent that writes social media posts matching our company tone"
See Creating & Configuring Agents for details.
Choosing the right agent
| What you want to do | Use this agent |
|---|---|
| General conversation, planning, coordination | Assistant |
| Web research, fact-checking, competitive analysis | Researcher |
| Write code, run commands, debug errors | Developer |
| Make phone calls, navigate phone menus | Receptionist |
| Something domain-specific | Create a custom agent |
TIP
You don't always need to choose manually. If you ask the Assistant something that needs research or coding, it can delegate to the Researcher or Developer automatically.
Next steps
- Creating & Configuring Agents. Build custom agents for your specific needs
- Delegation & Tasks. How agents work together