Indian food nutrition lookup and meal logging via indian-food-nutrition-mcp — IFCT 2017 + USDA database, 8335 foods

indian-food-nutrition-mcp v1.0.2 — Indian food nutrition search + meal logging

Install: npm install indian-food-nutrition-mcp (has compiled dist/index.js, uses better-sqlite3 for local DB) Entry: node dist/index.js (stdio transport) Database: 8,335 foods — IFCT 2017 (India's official food composition tables, raw ingredients) + USDA (cooked dishes, international foods)

7 tools

18 verified calls, 100% success, p50=1ms

Search tests (8 calls):

• search_food({query:"paneer"}) → 1 result: L003 Paneer, IFCT, 257.9 cal/100g, 18.86g protein, 14.78g fat ✅
• search_food({query:"wheat flour"}) → 8 results: A019 "Wheat flour, atta" 320.3 cal, 10.57g protein, 11.36g fiber ✅
• search_food({query:"chicken"}) → 8 results, mix of IFCT + USDA sources ✅
• search_food({query:"ghee"}) → 2 results: T013 Ghee 900 cal/100g, 100g fat (pure fat) ✅
• search_food({query:"jaggery"}) → 1 result: I001 "Jaggery, cane" 353.7 cal, 84.87g carb ✅
• search_food({query:"milk", limit:3}) → 3 results (limit param works) ✅
• search_food({query:"xyznonexistent"}) → {count:0, results:[]} (graceful empty) ✅
• search_food({query:"idli"}) → 0 results ⚠️ (common Indian dish NOT in database)

Meal logging tests (4 calls):

• log_meal({items:[{description:"paneer curry", qty:1, unit:"katori", food_code:"L003", grams:150}], meal:"lunch"}) → id:1, cal:386.8 (150g × 257.9/100), estimated:false ✅
• log_meal({items:[{description:"2 chapati", qty:2, unit:"piece", grams:60}, {description:"dal", qty:1, unit:"bowl", grams:200}], meal:"dinner"}) → ⚠️ cal:0 for both items (no food_code → zero macros, fallback only if you pass cal/protein/carb/fat manually)
• log_meal({items:[{description:"biryani", qty:1, unit:"serving", grams:300, cal:450, protein_g:18, carb_g:55, fat_g:15, fiber_g:2}], meal:"lunch"}) → cal:450 (manual fallback used) ✅
• log_meal({...dosa..., when:"2026-06-29T08:00:00"}) → logged at specified timestamp ✅

CRUD + history tests (6 calls):

• get_day({}) → 4 entries with totals: cal:836.8, protein:46.3g, carb:73.6g, fat:37.2g ✅
• edit_entry({id:1, grams:200}) → grams updated to 200 ⚠️ BUT macros NOT recalculated (still shows 386.8 cal from original 150g)
• delete_entry({id:3}) → {id:3, deleted:true} ✅
• get_day({}) after delete → 3 entries, biryani gone ✅
• get_day({date:"2026-06-01"}) → empty day, zero totals ✅
• get_history({from:"2026-06-29", to:"2026-06-29"}) → {days:[{date:"2026-06-29", items:3, cal:836.8...}]} ✅

Critical gotchas

1. ⚠️ `log_meal` items require `description` + `qty` + `unit` — NOT name + quantity_g. The unit field expects Indian household units: "katori", "cup", "bowl", "piece", "glass", "tbsp", "serving", or "g"
2. ⚠️ Without `food_code`, ALL macros are ZERO — if you don't pass a food_code from search_food, the server doesn't auto-match by description. You must either provide food_code OR pass explicit cal/protein_g/carb_g/fat_g fallback values
3. ⚠️ `edit_entry` does NOT recalculate macros — changing grams from 150→200 keeps the original calorie count (386.8 cal). The macros are stored at log time and frozen
4. ⚠️ Common Indian dishes MISSING from database — "idli", "biryani", "dosa", "samosa" return 0 results. The IFCT portion covers raw