← Sticker Wall AGENT · DETAIL

AI Diet Photo Check-in & Companion System: A Tool-Driven Private Domain System Generating 50k RMB/Month for a Solopreneur

Workflow: Every day, users send photos of three meals via WeChat Work. The workflow automatically invokes the vision large model t

AGENT

Key Fields

FIELD STAMPS

IndustryContent / Creator Economy

RegionChina

ScaleSME

ChannelOnline

🔧 Workflow

Every day, users send photos of three meals via WeChat Work. The workflow automatically invokes the vision large model to recognize meal types, portions, and cooking methods, matches them with the built-in Chinese calorie database to calculate total calorie intake and macronutrient ratios, and returns an analysis report and adjustment suggestions within 10 seconds. In the evening, after users upload their dinner photos, the system simultaneously generates the next day's all-day training and diet plan, automatically syncing it to the user's dedicated Lark (Feishu) spreadsheet for record-keeping. Abnormal data showing an intake exceeding 30% over the standard is automatically pushed to operations staff for manual follow-up.

🛠 Setup Requirements

First, register a Coze platform account, configure the image recognition node for the multimodal large model, and import a structured calorie database containing over 2,000 common Chinese dishes. Next, bind a WeChat Work customer service account, configure trigger rules for automatic message sending and receiving, and set up Lark Multi-dimensional Spreadsheets as the user data archive. No coding foundation is required overall; if familiar with low-code tool logic, the setup period is about 10 days, with an additional 3-5 days needed for testing and optimization.

🧰 Toolchain

🔧 Coze
🔧 WeChat Work
🔧 Doubao Vision Large Model
🔧 Lark Multi-dimensional Spreadsheet

💰 Revenue

① C-end fat loss and muscle gain companion subscription (main revenue): individual users subscribe and pay monthly, 499 RMB/month × about 100 companion clients = monthly revenue of about 50,000 RMB, accounting for about 100% of monthly revenue (figures derived from card data, self-reported by merchants, yet to be independently verified); ② Advanced customized muscle gain and preparation plans (periodic project service fee): paying users pay based on customized cycles with a unit price of 1,299 RMB, actual transaction volume is not public, and the proportion of total revenue is unknown (from case self-description, independent review still lacking); ③ Enterprise employee health management services (B-end subscription per enterprise): corporate HR pays for employee health management, neither unit price nor employee seat scale has been disclosed, the card estimates it can add 2,000-3,000 RMB in monthly revenue (merchant self-reported caliber, verification missing), and the exact share contributed by this part is unknown; ④ Opportunity: AI diet check-in SaaS self-service subscription, fitness users self-subscribe monthly, benchmarking against the EdgeOne Makers case in the card which launched 70 days ago, serving over 200 cumulative users, with a Monthly Recurring Revenue (MRR) of 28,000 RMB, proportion not disclosed.

💸 Cost

Vision interface call fees for large models cost about 1,800 RMB per month (calculated based on an average of 3 calls per user per day). The basic versions of WeChat Work and Lark Multi-dimensional Spreadsheets are free. If advanced automation features are used, an additional Coze platform membership fee of 99 RMB/month is required, bringing the overall monthly fixed cost to approximately 1,900 RMB.

⏱ Time Investment

2 hours per day (handling user abnormal diet data, community interaction Q&A, and customized plan adjustments), with an additional 3 hours required on weekends to review user weekly progress and iterate plans.

🚀 Getting Started

Step 1: Publish live-action content on Douyin and Xiaohongshu (RED) regarding AI diet check-ins and fat-loss meal calorie estimation to accumulate 1,000 targeted fitness fans; Step 2: Publish recruitment posts for companion services on platforms like Side Hustle Base (Fuyebase) and Jike, offering the first 10 trial users a discounted price of 199 RMB/month, and collect real-world usage scenario data to optimize model recognition accuracy; Step 3: Build a complete automated workflow, launch standardized services, and gradually raise prices to regular unit rates.

🔑 Keys to Success

✅ Recognition accuracy of the vision model for complex Chinese dishes, cooking methods, and portions, with errors controlled within 15%
✅ Operations staff possessing basic fitness and nutrition knowledge, capable of providing personalized adjustment plans tailored to user physical data and training goals
✅ Community operations providing emotional value, enhancing user retention and repurchase through daily check-in supervision and weekly progress reviews
✅ Continuous accumulation of user real dietary data to iterate the calorie database and improve model matching precision

⚠️ 风险

⚠️ High recognition errors by multimodal models for home-cooked and local specialty dishes (such as hot pot, spicy hot pot/malatang), which may lead to calorie estimation biases and trigger user trust crises
⚠️ Providing weight-loss guidance without being a registered dietitian may involve medical compliance risks, requiring clear disclaimer clauses in the service agreement
⚠️ If users experience health issues due to executing the plan, disputes may arise, necessitating the preliminary setup of accident insurance and professional nutritionist endorsement

📌 Real Cases

📌 In an independent fitness coach OEM case reported by Side Hustle Base, practitioners provided AI customized plan sample services for 12 personal trainers, achieving a single-month OEM income of 32,000 RMB
📌 In the AI fitness fat-loss diet guide shared by the Ti-xiao-lu platform, practitioners built an AI diet check-in companion system, serving 87 users in a single month and achieving a monthly income of 43,000 RMB
📌 In the EdgeOne Makers launch case shared by CSDN blogger Mrxiao_bo, their launched AI health coach service served over 200 cumulative users in 70 days, reaching a monthly recurring revenue of 28,000 RMB

Sources