On-premNDA

BOM Extractor

Feed it a single DXF drawing and a bill of materials (BOM) comes out within 15 seconds.

Precision-machining OEM · drawing-to-BOM extraction agent

Inside

How it works on the inside

Work that one person used to handle is split across multiple agents. The automation runs across the entire flow — not just a few steps.

01 · Agent

Drawing Agent

DXF parser + vision model auto-recognizes 118–216 drawing entities

02 · Agent

Vector DB Revision Map

Auto-maps revisions per part number with pgvector (JWED-12000: 3 revs)

03 · Agent

Material RAG

Searches the historical drawing parts catalog → auto-matches part name, material, and spec

04 · Agent

HITL Feedback Loop

Feeds expected-vs-actual result feedback back into the model

Stack

Tech stack

THROUGHPUT

8 drawings → 54-row BOM

STACK

DeepAgents · text-ai · pgvector · DXF Parser

STATUS

Demo Phase

On-prem deployment — an environment where DXF drawings can never leave the premises. Streamlit + self-hosted LLM. 8/8 DXF drawings → 54-row BOM auto-extracted (15.2 s/drawing) · automatic revision detection.
TreeSoop — Precision-machining OEM · drawing-to-BOM extraction agent

Other work in a similar vein

+300%

Sales AI Chatbot for a Top-5 Telecom Retailer

A chatbot trained on the tone and manner of veteran salespeople. Purchase conversion up 300% over the previous Kanana AI bot, absorbing the workload of 10+ CS and sales staff.

×5

Autopeople

AI agents split up the parts pricing and quote writing that used to run on Excel. Automatic price updates, auto-attached part photos, and margin optimization — a 5x quote close rate.

Facing a similar challenge?
Let’s talk it through

Excel workflows, chatbots, AI marketing — whatever the domain, it can be redesigned the same way.
The initial assessment and quote are free — the contact form is the fastest way to reach us.

Tell us about your project Browse more case studies