N8N Tools - Agno Knowledge Agent
Level 2 AI Agent with knowledge capabilities - Vector search, document indexing, and knowledge base powered by Agno framework
Overview
The N8N Tools - Agno Knowledge Agent node is an advanced AI agent designed to interact with knowledge bases using native large language model (LLM) capabilities. It supports vector search, document indexing, and semantic retrieval powered by the Agno framework. This node enables users to query a configured knowledge base with natural language instructions and messages, leveraging AI models from providers like OpenAI or Anthropic.
Typical use cases include:
- Semantic search across indexed documents for customer support or research.
- Automated knowledge retrieval and summarization from large text corpora.
- Combining keyword and vector search methods to improve information accuracy.
- Integrating AI-powered reasoning and external tools for complex workflows.
For example, a user can input a question about company policies, and the node will search the knowledge base, retrieve relevant documents, and generate a concise answer using the selected AI model.
Properties
| Name | Meaning |
|---|---|
| Knowledge Agent Recommendation | Informational notice describing the node's capabilities and setup instructions. |
| Model Provider | Select the AI model provider for processing knowledge queries: OpenAI or Anthropic (Claude). |
| Model | Choose the specific AI model based on the provider: OpenAI: GPT-4o, GPT-4o Mini, GPT-4 Turbo, GPT-3.5 Turbo Anthropic: Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3 Opus |
| Instructions | System prompt instructions guiding the AI assistant's behavior when interacting with the knowledge base. |
| Message | The input message or query string sent to the knowledge agent for processing. |
| Knowledge Base Configuration | Settings for the knowledge base: - Knowledge Base Name (identifier) - Search Method: Vector Search, Keyword Search, Hybrid Search - Max Results to return - Similarity Threshold for vector search - Enable Auto-Indexing of new documents - Chunk Size and Overlap for text indexing - Supported File Types (pdf, docx, txt, md, json, csv, html) - Enable Document Summarization |
| Enable Tools | Boolean to enable or disable native Agno tools integration. |
| Agno Native Tools | Select which native tools to enable for ultra-fast execution: Knowledge Tools, Reasoning Tools, Web Search Tools, Data Analysis Tools, Finance Tools, Email Tools, HTTP Request Tools, File System Tools, Database Tools, Shell Tools |
| Tool Configurations | Detailed configuration for enabled tools, including: - Knowledge Configuration (embedding model, chunk size/overlap) - File System Configuration (storage type, credentials) - Email Configuration (SMTP settings) - Database Configuration (type, connection string) - Finance Tools Configuration (default period, currency) |
| Advanced Options | Additional parameters: - Temperature (response randomness) - Max Tokens (max response length) - Output Format (Text, Markdown, JSON) - Enable Streaming (real-time responses) - Context Window Size for knowledge retrieval |
| Report Issue | Notice with links to report bugs or get help via support form or GitHub issues. |
Output
The node outputs a JSON object per input item containing:
response: The AI-generated response text from the knowledge agent.agent_type: Fixed value"knowledgeAgent".agent_level: Fixed value2indicating the agent sophistication level.execution_time_ms: Time taken for the agent to process the request in milliseconds.knowledge_base_searches: Number of knowledge base searches performed.documents_retrieved: Count of documents retrieved from the knowledge base.search_results_count: Number of search results returned.similarity_scores: Array of similarity scores for retrieved documents.tools_used: List of native tools utilized during processing.llm_calls: Number of LLM calls made.llm_tokens_used: Token usage statistics for the LLM.cache_hits: Number of cache hits during processing.session_id: Unique session identifier for the request.agno_version: Version of the Agno framework used.model_used: The AI model name selected.model_provider: The AI model provider selected.knowledge_base_used: Name of the knowledge base queried.search_method_used: Search method applied (vector, keyword, hybrid).input_data: Original input JSON data if provided.
If an error occurs, the output JSON contains an error field with the error message and success set to false.
The node does not output binary data.
Dependencies
- Requires an API key credential for the N8N Tools API to authenticate requests.
- Connects to the Agno backend service at
https://n8ntools-agno-production.up.railway.app/api/v1/agent. - Supports integration with various external services depending on enabled tools, such as:
- Cloud storage (Amazon S3, Cloudflare R2, MinIO, Google Cloud Storage)
- SMTP email servers
- SQL and NoSQL databases (PostgreSQL, MySQL, SQLite, MongoDB)
- Financial data sources (YFinance)
- Proper configuration of these external services and credentials is required for full functionality.
Troubleshooting
- Missing API Credentials: The node requires valid API credentials; ensure the API key is configured correctly in n8n credentials.
- Network Errors: Connectivity issues to the Agno API endpoint may cause failures; verify network access and endpoint availability.
- Invalid Model or Provider Selection: Selecting unsupported or mismatched model-provider combinations can cause errors.
- Exceeding Token Limits: Responses exceeding max tokens or context window sizes may be truncated or fail; adjust advanced options accordingly.
- Incorrect Knowledge Base Configuration: Misconfigured search methods or knowledge base names may result in no results or errors.
- Error Messages: The node returns error messages from the Agno API in the output JSON under the
errorfield. Review these messages for troubleshooting. - Continue On Fail: If enabled, the node will continue processing subsequent items even if some fail, returning error details per item.