Overview
This node implements an ultra-fast AI agent powered by the Agno framework, designed for advanced multi-agent systems with microsecond-level performance. It supports multiple agent types ranging from a basic AI assistant to complex team-based and workflow-driven agents. The node is useful for scenarios requiring intelligent automation, reasoning, knowledge retrieval, collaborative problem solving, or orchestrated workflows.
Practical examples include:
- A basic conversational AI assistant that answers queries quickly.
- A knowledge agent that integrates with databases or knowledge bases.
- A reasoning agent capable of advanced logic and memory.
- A team agent coordinating multiple specialized AI roles (e.g., data analyst, content writer).
- A workflow agent managing stateful, conditional, or iterative processes.
The node accepts input messages and instructions, optionally enables native tools for enhanced capabilities, and can maintain conversation memory for context persistence.
Properties
| Name | Meaning |
|---|---|
| Agent Type | Selects the type of AI agent: - ⥠Basic Agent (Level 1): Fast execution with tools and instructions - ð Knowledge Agent (Level 2): Includes knowledge base and storage - ð§Ū Reasoning Agent (Level 3): Memory and advanced reasoning - ðĨ Team Agent (Level 4): Multi-agent collaboration - ð Workflow Agent (Level 5): Workflow automation with state management |
| Instructions | System instructions guiding the agent's behavior (e.g., "You are a helpful AI assistant..."). |
| Message | Input message or query sent to the agent. |
| Enable Tools | Boolean to enable or disable Agno native tools for the agent. |
| Agno Native Tools | Select which native tools to enable when tools are enabled: - ð§Ū Reasoning Tools - ð Web Search Tools - ð Data Analysis Tools - ð° Finance Tools (YFinance) - ð§ Email Tools - ð HTTP Request Tools - ð File System Tools - ðïļ Database Tools - ð§ Knowledge Tools - ð Shell Tools |
| Tool Configurations | Configuration options for enabled tools, such as database connection details, email SMTP settings, file system storage provider and credentials, HTTP request defaults, finance tool parameters, and shell command settings. |
| Advanced Options | Additional settings including: - Temperature: Controls randomness in responses (0 deterministic to 2 very random) - Max Tokens: Maximum tokens in response - Enable Memory: Enable conversation memory and context persistence - Memory Configuration: Session ID, memory type (conversation buffer, summary, window, token buffer), max messages/tokens, persistent storage options (in-memory, PostgreSQL, MongoDB, Redis, SQLite) - Enable Streaming: Enable streaming responses - Multi-Modal Input: Enable image/audio/video input processing - Output Format: Text, JSON, Markdown, or HTML - Fallback Model: Alternative model if primary fails |
| Team Configuration | (Visible only if Agent Type is Team Agent) Settings for team size, collaboration strategy (sequential, parallel, hierarchical, democratic), and specialist roles (data analyst, content writer, researcher, designer, developer, mathematician, strategist, problem solver). |
| Workflow Configuration | (Visible only if Agent Type is Workflow Agent) Settings for workflow type (linear, branching, loop, state machine), max iterations, and checkpoint enabling for saving workflow state. |
Output
The node outputs a JSON object per input item containing:
response: The main textual or structured response from the agent.usage: Information about resource usage (e.g., tokens consumed).model: The language model used for generating the response.agentType: The selected agent type.executionTime: Time taken for the agent to process the request.reasoning: Optional detailed reasoning or chain-of-thought data if available.toolsUsed: List of tools utilized during the agent's execution.memoryKey: Identifier for the memory session if memory is enabled.- If streaming is enabled, includes
streaming: trueand astreamUrlfor real-time response consumption.
The node does not output binary data.
Dependencies
- Requires connection to an AI language model node providing the model input.
- Requires an API key credential for the N8N Tools API service (Agno backend).
- Optional dependencies based on enabled tools:
- Database connections (PostgreSQL, MySQL, SQLite, MongoDB) require valid connection strings.
- Email tools require SMTP server credentials.
- File system tools require cloud storage credentials (Amazon S3, MinIO, Google Cloud Storage, Cloudflare R2).
- HTTP request tools use configurable headers and SSL verification.
- Finance tools rely on YFinance data.
- Shell tools execute commands with configurable shell environment.
- Memory persistence may require additional database or cache services (PostgreSQL, MongoDB, Redis, SQLite) depending on configuration.
Troubleshooting
- Missing AI Language Model Input: Error "AI Language Model input is required" occurs if the node is not connected to a compatible AI model node. Ensure the model node is connected to the first input.
- Missing API Credentials: Error "N8N Tools API credentials are required" indicates the API key credential is not set or invalid. Configure the required API key credential properly.
- Tool Configuration Errors: Misconfigured database connection strings, SMTP credentials, or cloud storage keys can cause failures. Verify all tool-specific configurations.
- Timeouts or Slow Responses: Adjust timeout and retry settings in advanced options or tool configurations.
- Memory Persistence Issues: If using external storage for memory, ensure connectivity and correct credentials for the chosen database/cache.
- Streaming Not Working: Streaming requires version >=1.1 and must be enabled explicitly; check compatibility and network access to stream URL.
- Unsupported Agent Type Features: Some properties like teamConfig or workflowConfig apply only to specific agent types; setting them incorrectly may have no effect.