You are helping me build the following AI agent workflow.

## Goal
Telegram Nutrition Tracker & Meal Logger: Log meals via text or image in Telegram to auto-calculate macros and track goals against a daily progress report.

## Specification
- What it does: Log your meals instantly via text, voice messages, or photos directly from Telegram without needing a dedicated app. Automatically calculate calories and macros while tracking your personalized nutrition goals in real time. Receive daily progress summaries with visual updates sent straight to your chat.
- Trigger: Triggered by an event (Event · on new Telegram message)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - Telegram: receiving user inputs like text, voice, or images
  - Google Gemini Chat Model: analyzing food images and transcribing voice messages
  - Google Sheets: storing user profiles and meal logs
  - Code: processing data for daily reports

## Known pitfalls, handle each one explicitly in your implementation
1. Image recognition often fails on blurry or dark photos, so add a validation step that asks users to retake the picture if confidence is low.
2. Nutrition estimates vary significantly between different food brands and cooking methods, so include a disclaimer in every summary about potential inaccuracies.
3. Google Sheets quotas can be exceeded by frequent writes from high-volume bots, so batch updates into hourly summaries instead of real-time entries.
4. Voice-to-text transcripts frequently mishear specific food names like 'oatmeal' or 'broccoli', so run a secondary spell-check pass before calculating macros.

## Reference implementation
https://n8n.io/workflows/7756 (workflow template)
Fetch and inspect this before building. If it matches my stack, adapt it;
if not, rebuild the pattern with my tools.

## Process requirements
1. Before building: ask me which of the listed tools I actually use and
   what my platform is (n8n / Make / code / other). Do not assume.
2. Adapt the pattern to my answers; do not force the reference stack.
3. Address every pitfall above; tell me how you handled each.
4. Provide a test plan I can run before letting this touch real data.
5. Ask before any step that sends messages, modifies data, or spends money.

Source: https://usecasesforagents.com/use-case/telegram-nutrition-tracker-meal-logger/ via usecasesforagents.com