mcp-talent-server/dist/tools/vector-search.d.ts

Model Context Protocol server for talent management tools

import { z } from 'zod';
declare const VectorSearchInput: z.ZodObject<{
    query: z.ZodString;
    limit: z.ZodDefault<z.ZodNumber>;
}, "strip", z.ZodTypeAny, {
    limit: number;
    query: string;
}, {
    query: string;
    limit?: number | undefined;
}>;
type VectorSearchInputType = z.infer<typeof VectorSearchInput>;
export declare class VectorSearchTool {
    search(input: VectorSearchInputType): Promise<any>;
}
export declare const VectorSearchDescription = "**Vector Search Tool - Semantic Document & Knowledge Base Search**\n\n**Primary Purpose**: Advanced semantic search across uploaded documents and talent rate cards using AI-powered embeddings.\n\n**Core Capabilities**:\n\u2022 **Semantic Understanding**: Finds conceptually related content, not just keyword matches\n\u2022 **Multi-Document Search**: Searches across PDF, DOCX, TXT files simultaneously\n\u2022 **Relevance Scoring**: Returns results with confidence scores (0-1 scale)\n\u2022 **Content Highlighting**: Provides contextual snippets with key terms highlighted\n\u2022 **Smart Filtering**: Filter by file type, user, relevance threshold, and date\n\u2022 **Deduplication**: Groups results by document to avoid redundant chunks\n\u2022 **Rich Metadata**: Includes document source, creation date, file type, and processing stats\n\n**Specialized for Talent Management**:\n\u2022 **Rate Card Search**: Find specific pricing for Instagram Stories, TikTok posts, YouTube integrations\n\u2022 **Talent Discovery**: Search for influencers by niche, audience size, or platform specialization\n\u2022 **Contract Information**: Locate terms, deliverables, and pricing structures\n\u2022 **Performance Data**: Find engagement rates, reach statistics, and campaign results\n\n**Advanced Features**:\n\u2022 **Multi-criteria Filtering**: Combine relevance, date, file type, and user filters\n\u2022 **Batch Processing**: Handle large document collections efficiently\n\u2022 **Error Recovery**: Robust error handling with helpful suggestions\n\u2022 **Search Analytics**: Provides processing time, match counts, and quality metrics\n\n**Optimal Use Cases**:\n\u2713 \"Find Instagram Story rates for beauty influencers\"\n\u2713 \"Show me YouTube integration pricing above $5000\"\n\u2713 \"Search for talent contracts with exclusive clauses\"\n\u2713 \"Find performance data for TikTok campaigns in Q3\"\n\u2713 \"Locate audience demographics for fitness influencers\"\n\n**Important Note**:\n- For searching for rates always search in Rate Sheet 2025 file vectors for accurate results.\n- Always mention file name in the query to get accurate results.\n\nexample vector search queries:\n- Denny kim rates in Rate Sheet 2025\n\n**Parameters Guide**:\n\u2022 query: Natural language search (required)\n\u2022 limit: Number of results to return (default: 10, max: 100)\n\n**Response Format**: Returns structured results with content snippets, relevance scores, source attribution, search statistics, and actionable suggestions for query refinement.\n\n**Important Note**:\n\u2022 This tool use pinecone vector store always use limit parameter to paginate the results and optimize the results and output tokens";
export declare const vectorSearchSchema: z.ZodObject<{
    query: z.ZodString;
    limit: z.ZodDefault<z.ZodNumber>;
}, "strip", z.ZodTypeAny, {
    limit: number;
    query: string;
}, {
    query: string;
    limit?: number | undefined;
}>;
export {};
//# sourceMappingURL=vector-search.d.ts.map