# breathstrap-sourcing-agents

## Status
wip — initial research underway

## Primary goal
Find **China-based sourcing agents, supply-chain consultants, procurement scouts, and hardware-startup sourcing people** who are likely capable of finding real pneumatic blood-pressure-watch ODMs. We are **not** looking for the factories themselves in this task — only the people who can find and qualify those factories.

## Why this matters (context for the researcher)
- **Product:** a non-medical pneumatic wrist wearable (codename: *breathstrap*).
- **Mechanism:** a soft air bladder, micro pump, valves, and pressure control produce a gentle rhythmic squeeze/release cue on the wrist.
- **Target supplier class:** factories that already make real air-pump + airbag blood-pressure watches or wrist blood-pressure monitors (i.e., the closest existing ODM/ODM-adjacent capability to what breathstrap needs).
- **Two-step funnel:**
  1. **This project** → identify sourcing agents / scouts who *find* such factories.
  2. (later) → use those agents to identify, visit, and qualify the actual factories.
- We explicitly do **not** want generic "we source anything" agents. We want people with domain fit in **wearables, medical electronics, micro-pump/airbag systems, or hardware-startup ODM work**.

## Geography
Primary: **Shenzhen / 珠三角 / 华强北** ecosystem, with secondary reach to **Dongguan, Suzhou, Shanghai, Beijing, Xiamen** where relevant supply clusters exist.

## Research approach
- **Primary discovery:** [Sogou Weixin](https://weixin.sogou.com/) (微信公众号 articles — the dominant long-form / professional-sourcing content surface in China).
- **Supplementary search:** Baidu, 企查查 (Qichacha), 天眼查 (Tianyancha), LinkedIn, company websites, public registries.
- **Why supplement Sogou:** Sogou's WeChat index is incomplete — it misses articles older than ~1 year, does not index every public account, often strips author/contact info, and can return stale or duplicate snippets. We document what it misses and use Baidu + general web search to fill the gaps.

## Search terms (this run)
```
深圳 寻源 顾问 智能硬件
深圳 采购代理 智能穿戴
深圳 供应链顾问 智能硬件
深圳 医疗电子 采购代理
深圳 智能穿戴 供应链
深圳 智能手表 ODM 寻源
气泵血压手表 供应链
气囊血压手表 ODM 资源
腕式血压计 供应链
医疗电子 寻源 深圳
华强北 采购顾问 智能硬件
硬件创业 深圳 采购 顾问
```

## Per-result workflow
For each promising result:
1. Open the WeChat article.
2. Extract author / account name.
3. Capture: WeChat ID, phone, QR code, company name, website, location, service description.
4. Record evidence of: sourcing, supplier verification, factory visits, ODM/OEM work, electronics, wearables, medical electronics, or hardware-startup support.
5. Cross-check the company / person on Baidu, 企查查, 天眼查, LinkedIn, company website, or other public pages.
6. Score 1–5 on fit (5 = strong domain fit + verifiable identity + active recent work).

## Files in this project
- `README.md` — this file.
- `log.md` — chronological research log (queries run, links opened, observations, dead ends).
- `candidates.csv` — structured candidate list (one row per lead).

## Scoring rubric (1–5)
- **5 — Strong lead:** named individual or company, domain-verified (wearables / medical electronics / pump+airbag), public contact, active recent (≤12 mo) sourcing or factory-intro work.
- **4 — Promising:** domain fit clear, identity verifiable, contact present but maybe not all channels.
- **3 — Possible:** some fit, but identity or domain thin; needs follow-up.
- **2 — Weak:** generic sourcing agent, no clear wearable/medical electronics focus.
- **1 — Discard:** no sourcing/consulting signal, wrong domain, or unverifiable.

## Output / hand-off
A shortlist of 5–15 candidates, ranked, with contact info and the evidence behind each score. The downstream task (separate project) will be to use these agents to find and qualify the actual pneumatic BP-watch ODMs.