ProjectIntent Over Keywords: From Stitched Embeddings to a Production Fashion Agent · Module 15: Clarification policy and the cap · part 17 of 26
Part 17 · Module 15: Clarification policy and the cap

Module 15: Clarification policy and the cap

24 min read·
About 45 minutes · Needs: Module 13 (sessions), Module 14 (the planner) · Code: part3-styleme/backend/src/styleme/agent/planner.py, agent/prompts.py, agent/parsing.py, sessions.py, api.py, part3-styleme/frontend/src/lib/clarify.ts

Where this fits

WHERE THIS FITS Intent Over Keywords this module PART 1: STITCHED VIDEO RAG three models, two spaces, RRF PART 2: UNIFIED VIDEO RAG one model, one space PART 3: STYLEME FastAPI, LangGraph, React QUALITY AND OPERATIONS evals, Docker, Cloud Run Frames ffmpeg, one per 5 s CLIP ViT-B/32 frames + text, 512 d Whisper base speech to text MiniLM 64-word chunks, 384 d Two collections p1_visual, p1_text Rank fusion RRF, k = 60 Gemini answer from transcript text API + Streamlit upload, ask, sources Windows ffmpeg, video + audio Embedding 2 Gemini, one vector One collection p2_unified, no fusion Notes + answer Gemini, cited times React web chat, photo, chips nginx one URL, /api proxy FastAPI /api/chat, limits Sessions in memory, cap of 3 Planner intent before retrieval Implementer batched embed, search Generator stylist reply, ids Grounding check no LLM, drops invented Gemini flash-lite, embedding 2 Qdrant styleme_products Catalogue ingest image + title vectors Langfuse optional traces Test video 26 questions, 4 kinds Video eval Part 1 vs Part 2 Agent evals intent, grounding Guards uploads, limits, secrets Containers compose profiles Cloud Run StyleMe only CI/CD GitHub Actions, WIF Original code v1, v2, fashion-agent
This module (highlighted): the planner's rule for asking, the session that counts the questions, the API that updates the count and the web page that shows it. Blue is our own code.

Module 14 showed how the planner picks an intent before any retrieval runs. This module zooms in on one intent: the clarification question. When should StyleMe ask instead of search, how many times may it ask, who counts, and what happens when the model wants one question more than allowed? The answers live in the planner, the session that holds the count, the API that updates it and the page that shows "Question 2 of 3".

The rest of this project is yours to keep

This project is bought on its own, once, and stays readable afterwards, including the parts added to it later.

Buy this project · ₹799See what is inside

Or open everything with a plan