Agent Library
Create, manage, and run AI agents for automated tasks. Agents are specialized AI personas that excel at specific roles in your workflows.
What are AI Agents?#
AI Agents in PromptReports are specialized AI personas configured for specific tasks. Each agent has a defined role, system prompt, capabilities, and model configuration. Agents can work individually or together in multi-agent workflows to tackle complex report generation tasks.
Specialized Personas
Each agent is optimized for a specific type of task.
Multi-Agent Workflows
Combine agents for complex, multi-step processes.
Full Configuration
Control model, temperature, and capabilities.
Automated Execution
Agents can run autonomously on schedules.
System vs Custom Agents
Agent Types#
The Agent Library supports two types of agents:
| Type | Description | Editable |
|---|---|---|
| System Agents | Pre-built agents optimized for common tasks | View only (not editable) |
| Custom Agents | User-created agents for specific needs | Fully editable |
System Agents
Battle-tested agents ready for immediate use.
Custom Agents
Create agents perfectly suited to your workflow.
Agent Roles#
Agents are categorized by their primary role, which determines their strengths:
| Role | Description | Example Tasks |
|---|---|---|
| Researcher | Gathers and synthesizes information | Market research, competitive analysis, data gathering |
| Analyst | Processes data and extracts insights | Trend analysis, statistical review, pattern recognition |
| Writer | Creates polished, engaging content | Report drafting, executive summaries, blog posts |
| Reviewer | Validates quality and accuracy | Fact-checking, consistency review, quality assurance |
| Editor | Refines and improves content | Grammar, tone adjustment, structural improvements |
| Specialist | Domain-specific expertise | Technical review, legal compliance, industry analysis |
Creating Agents#
Create custom agents tailored to your specific needs:
Click "Create Agent"
Name and Role
System Prompt
Capabilities
Model Settings
Save
System Prompt Best Practices
- The agent's expertise and background
- Their communication style and tone
- Specific instructions for their role
- What they should and shouldn't do
- How they should format their outputs
Agent Configuration#
Each agent can be configured with these settings:
| Setting | Description | Recommendation |
|---|---|---|
| Model | Which AI model powers the agent | GPT-4 for complex tasks, GPT-3.5 for simpler ones |
| Temperature | Creativity vs consistency (0-1) | 0.3-0.5 for analysis, 0.7-0.9 for creative writing |
| System Prompt | Core instructions and personality | Be specific and comprehensive |
| Capabilities | Tags describing agent abilities | Use for filtering and workflow matching |
Search Agents
Find agents by name or capability.
Filter by Role
Narrow down to specific agent roles.
Filter by Type
Show system or custom agents only.
View Details
Click any agent to see full configuration.
Pro Tips#
Best Practices
- Start with system agents to understand the patterns before creating custom ones
- Keep agents focused - one agent should excel at one type of task
- Use capability tags consistently to make agents easier to find and match
- Test agents thoroughly before using them in production workflows
- Lower temperature for analytical tasks, higher for creative tasks
- Review agent outputs regularly and refine system prompts based on results
- Build a library of agents that complement each other in workflows
Agents are the building blocks of automated content generation. A well-designed agent library allows you to tackle any report generation task by combining the right specialists. See the Workflows documentation to learn how to chain agents together.