Log meals by photo, barcode, or one tap. Then connect ChatGPT or Claude — your AI can read your nutrition and even set your macro plan.
FeedAI uses the Model Context Protocol (MCP) to connect your nutrition data — plus sleep, recovery, body composition, and activity from your connected health services — to ChatGPT, Claude, and other AI assistants. New in 1.3: they don't just read — they can set your daily macro plan.
ChatGPT, Claude, and any MCP-compatible assistant can read your nutrition and wearable data — and write your macro plan. Ask questions wherever you're already working.
Snap a photo, tap save. AI estimates macros in seconds. No more searching databases or weighing portions.
Grant and revoke access to your data anytime. Your food logs are yours - we just make them useful.
Snap a photo, scan a barcode, or type "two eggs and toast." AI estimates macros with a confidence score. Not perfect, but fast.
Most meals are repeats. Pin your regulars, or tap one from "Suggested now" — your usuals, ranked by when and where you eat them. Wrong meal? 5-second Undo.
Ask ChatGPT or Claude about your nutrition — or let them set your macro plan. Your rings update to match.
Simple interface, powerful data behind the scenes
From quantified-self enthusiasts to busy professionals who just want answers.
Simple tools that create useful data
Snap a photo and AI identifies the food. Scan packaged stuff against OpenFoodFacts and USDA data. Or just type "two eggs and toast." You might want to adjust portions sometimes.
Most meals are repeats, so the second time is one tap. Pin your regulars, relog anything from history, and "Suggested now" surfaces your usuals by time of day and place.
Pick 3-8 dashboard rings from 17 nutrients. Goals fill up as you eat; limits are budgets that shift green to amber to red as you approach FDA/AHA caps. Targets come pre-filled — zero setup. Tap any ring to see which foods drove it.
Your rings on your Home Screen — top 4 in a small square, or every ring you track in the wide widget. Both update live as you log.
ChatGPT and Claude can read your nutrition, sleep, and recovery. Your AI coach can also set your macro plan — carb cycling, race-week fueling — with the rationale shown in-app.
Every meal is an ingredient list, not a blob. Fix one line — "actually 4 eggs, not 2" — and re-analyze with AI to update the numbers instantly.
Free to use every day. Premium unlocks the heavy AI lifting — 30-day free trial, no credit card required.
Model Context Protocol is an open standard that lets AI assistants access external data. FeedAI implements MCP so ChatGPT, Claude, and other compatible tools can read your nutrition data — plus sleep, recovery, body composition, and activity from your connected health services. They can also write: set your daily macro plan and update your goals as you chat.
It's good for common foods and reasonable for most meals. The AI gives a confidence score and asks a clarifying question when it's unsure. Think of it as a solid first guess that saves you time - you can always adjust before saving. Not perfect, but fast.
Nope. The app works great on its own - photo logging, one-tap repeats, rings, and widgets don't need any AI connected. MCP integration is there if you want it.
Yes. You control which AI assistants can access your data, and you can revoke access anytime. FeedAI never sells your data.
Most rings fill up as you eat - protein, fiber, potassium. Budget rings work the other way: they're limits. Added sugar, sodium, and saturated fat shift from green to amber to red as you approach FDA/AHA caps. Added sugar uses the AHA limit — 25g for women, 36g for men — and natural fruit sugar doesn't count against it. Targets come pre-filled - long-press any ring to customize.
Yes, really. Connected assistants can write to FeedAI, not just read. Tell Claude you're tapering for a marathon and it sets your carbs for each day of race week - your rings update to match, and the rationale shows up in the app. You can edit or remove anything it sets, anytime.
FeedAI