@radzor/ai-classifier

Classify text into categories using LLM-powered classification or rule-based approaches. Supports single and batch classification, custom category definitions with descriptions, and a trainable keyword-based fallback for offline use.

AI & MLv0.1.0typescriptpythonServerclassifierclassificationnlptextaillmcategorizationintentby Radzor

Install

View source on GitHub →

$npx radzor@latest add ai-classifier

⚠ Constraints: LLM mode requires an OpenAI-compatible API key. Rules mode works offline with no API key. Hybrid mode tries rules first and falls back to LLM when rule confidence is low.

Inputs

Name	Type	Default	Description
apiKey	string	—	API key for the LLM used for classification (OpenAI-compatible).OPENAI_API_KEY
model	string	gpt-4o-mini	Model to use for LLM-based classification.
baseUrl	string	https://api.openai.com/v1	Base URL for the classification API.
mode	'llm' \| 'rules' \| 'hybrid'	llm	Classification mode: llm (API-based), rules (keyword-based), or hybrid (rules first, LLM fallback).

apiKeystring

API key for the LLM used for classification (OpenAI-compatible).

OPENAI_API_KEY

{ "llm": { "constraints": "LLM mode requires an OpenAI-compatible API key. Rules mode works offline with no API key. Hybrid mode tries rules first and falls back to LLM when rule confidence is low.", "usageExamples": "llm/examples.md", "integrationPrompt": "llm/integration.md" }, "name": "@radzor/ai-classifier", "tags": [ "classifier", "classification", "nlp", "text", "ai", "llm", "categorization", "intent" ], "events": [ { "name": "onClassified", "payload": { "text": "string", "method": "string", "category": "string", "confidence": "number" }, "description": "Fired after each successful classification." } ], "inputs": [ { "name": "apiKey", "type": "string", "envVar": "OPENAI_API_KEY", "required": false, "description": "API key for the LLM used for classification (OpenAI-compatible)." }, { "name": "model", "type": "string", "default": "gpt-4o-mini", "required": false, "description": "Model to use for LLM-based classification." }, { "name": "baseUrl", "type": "string", "default": "https://api.openai.com/v1", "required": false, "description": "Base URL for the classification API." }, { "name": "mode", "type": "'llm' | 'rules' | 'hybrid'", "default": "llm", "required": false, "description": "Classification mode: llm (API-based), rules (keyword-based), or hybrid (rules first, LLM fallback)." } ], "radzor": "1.0.0", "actions": [ { "name": "classify", "params": [ { "name": "text", "type": "string", "description": "The text to classify." } ], "returns": "Promise<ClassificationResult>", "description": "Classify a single text into one of the defined categories." }, { "name": "batchClassify", "params": [ { "name": "texts", "type": "string[]", "description": "Array of texts to classify." } ], "returns": "Promise<ClassificationResult[]>", "description": "Classify multiple texts. Processes sequentially for LLM mode, in parallel for rules mode." }, { "name": "defineCategory", "params": [ { "name": "name", "type": "string", "description": "Category label." }, { "name": "description", "type": "string", "description": "Description of what belongs in this category (used by LLM classifier)." }, { "name": "keywords", "type": "string[]", "required": false, "description": "Keywords for rule-based matching." } ], "returns": "void", "description": "Define a classification category with its description and optional keywords." }, { "name": "train", "params": [ { "name": "examples", "type": "Array<{ text: string; category: string }>", "description": "Labeled examples to extract keywords from for rule-based classification." } ], "returns": "{ categoriesUpdated: number; keywordsAdded: number }", "description": "Train the rule-based classifier by extracting keywords from labeled examples." } ], "outputs": [ { "name": "classificationResult", "type": "ClassificationResult", "fields": [ { "name": "category", "type": "string", "description": "The predicted category label." }, { "name": "confidence", "type": "number", "description": "Confidence score for the top category (0-1)." }, { "name": "scores", "type": "Record<string, number>", "description": "Confidence scores for all categories." }, { "name": "method", "type": "string", "description": "The method used: 'llm' or 'rules'." } ], "description": "The classification result with category, confidence, and per-category scores." } ], "runtime": "server", "version": "0.1.0", "category": "ai", "languages": [ "typescript", "python" ], "description": "Classify text into categories using LLM-powered classification or rule-based approaches. Supports single and batch classification, custom category definitions with descriptions, and a trainable keyword-based fallback for offline use.", "dependencies": { "radzor": [], "packages": {} }, "composability": { "connectsTo": [ { "output": "classificationResult", "mapField": "category", "compatibleWith": [ "@radzor/event-tracker.action.track.eventName", "@radzor/log-aggregator.action.info.message" ] }, { "event": "onClassified", "description": "Route classified content to appropriate handler", "compatibleWith": [ "@radzor/agent-router.action.route.prompt", "@radzor/notification-hub.action.send.body" ] } ] } }

@radzor/ai-classifier

Install

Inputs

Outputs

Actions

Events

Composability

radzor.manifest.json