Kodexo Rag

Search Kodexo Labs Knowledge Base

Search the Kodexo Labs internal knowledge base for relevant document chunks using vector similarity search. Returns ranked document chunks matching the query, with similarity scores and source metadata. Use this for finding specific information, facts, or content from the 21-document knowledge base. Categories: A (Shared Foundation — brand, voice, team), B (Content Team — services, case studies, SEO), C (Lead Gen & Outreach — ICP, cold email, pricing), D (Community), E (Sales & Conversion), F

Ask Kodexo Labs Knowledge Base

Ask a question and get an AI-generated answer grounded in the Kodexo Labs knowledge base with source citations. Uses GPT-4o to generate a factual, grounded answer based ONLY on the internal knowledge base content. The answer includes source document references. Use this when you need a synthesized answer rather than raw document chunks. This tool takes longer than search (5-15 seconds) because it runs: vector search → context injection → GPT-4o generation. Categories: A (Shared Foundation), B

List Knowledge Base Documents

List all documents in the Kodexo Labs knowledge base with optional filters. Returns a paginated list of documents with their metadata (doc_code, title, category, status, version, chunk count). Use to browse available documents or find a specific one. Args: - category_code (string): Filter by category A-F or empty for all - status (string): Filter by 'active', 'draft', 'archived', or empty for all - search (string): Search by title or doc_code - page (number): Page number (default: 1)

Get Knowledge Base Document

Get full details of a specific document from the knowledge base by doc_code. Returns document metadata, and optionally the full content and individual chunks. Use kodexo_list_documents first to find available doc_codes. Args: - doc_code (string): Document code (e.g., 'A1', 'B3') - include_chunks (boolean): Include individual chunks (default: false) - include_content (boolean): Include full raw content (default: false) Returns: Document details with metadata, and optionally content/chunk

Upload Document to Knowledge Base

Upload a new document or update an existing one in the knowledge base. The document content is automatically chunked (heading-aware, 800-1500 chars), embedded via OpenAI, and stored in the vector database. If a doc_code already exists, it creates a new version (old version is preserved in history). WARNING: This operation can take 30-120 seconds for large documents due to embedding. The first letter of doc_code determines the category (A→Shared Foundation, B→Content Team, etc.). Args: - doc

List Knowledge Base Categories

List all 6 knowledge base categories with document counts. The knowledge base is organized into 6 categories: - A: Shared Foundation (brand identity, voice guidelines, team info, competitors) - B: Content Team (services, case studies, SEO keywords, templates) - C: Lead Gen & Outreach (ICP, cold email, objection handling, pricing) - D: Community (engagement playbook) - E: Sales & Conversion (discovery calls, testimonials) - F: Operations (client onboarding SOPs) No arguments required. Returns: