Meals in plain language
Say “oatmeal with banana and peanut butter” — your AI estimates calories and macros, fiber, total sugar and caffeine included, and logs it. No database search.
Lose It! alternative
Lose It! doesn't have one — so you can't use it inside Claude or ChatGPT. Nutrition MCP does the same job by conversation, and it's free and open source.
The short answer
The Model Context Protocol (MCP) is the open standard that lets AI assistants like Claude and ChatGPT connect to outside tools. Lose It! doesn't publish an MCP server, so there's no official way to log food to it from your AI. If you searched for “Lose It! MCP” or “connect Lose It! to Claude,” what you're really after is a nutrition tracker that lives inside your AI — that's exactly what Nutrition MCP is.
What you get instead
Say “oatmeal with banana and peanut butter” — your AI estimates calories and macros, fiber, total sugar and caffeine included, and logs it. No database search.
Send a product barcode and pull the label macros from Open Food Facts — fiber and sugar too, where the label lists them. No Premium subscription to unlock it.
Log body weight in kg or lb, set calorie, macro, fiber, sugar, caffeine, and water goals — fiber a target to reach, sugar and caffeine limits to stay under — and track trends toward a goal weight. Alcohol tracking is there too, opt-in and off unless you turn it on.
Ask for daily totals, weekly trends, streaks, and recurring meal patterns — right in the chat.
Import your history from another app's CSV export — parsed in your browser, not by the AI. Export it all back out, or delete your account, whenever you want.
MIT-licensed and self-hostable — no ads, no paywall, no upsell. Audit the code or run your own instance.
Lose It! vs. Nutrition MCP
Lose It! is a friendly calorie counter. Nutrition MCP does the same core logging by conversation, free, without ever leaving Claude or ChatGPT.
Moving from Lose It!
Lose It! won people over by keeping calorie counting light and a little gamified, with its Snap It photo logging as the headline trick. Nutrition MCP does the photo trick too — send a picture of your plate and your AI reads it — except it lives inside the assistant you already chat with, so there's no separate app to open.
If what you liked about Lose It! was low-friction logging and quick daily feedback, you'll feel at home: say what you ate, get your remaining calories and macros back, and move on. No ads, no upsell, and no account to juggle.
The one thing you give up is the streaks-and-badges layer Lose It! uses to keep you coming back. If that gamification is what motivates you, that's a fair reason to stay. If it always felt like noise on top of the actual logging, you won't miss it — the daily number is right there in the chat whenever you ask.
Your Lose It! history
Ask to import and an importer opens right in the chat: pick your export, map the columns, preview what will be added, then confirm. The file is read in your browser — the AI never sees the rows. In clients without in-chat panels, paste your export instead.
Switching doesn't mean starting from zero. Ask to import and an importer opens in the chat: you pick the CSV Lose It! exports, it's parsed in your browser, the columns it recognises map themselves — the date, food, meal, calories, protein, carbs, and fat, plus fiber, total sugar, and caffeine where your export carries them — and you confirm a preview of what will be added. It's a file picker and a preview, not a dictation exercise — on that path the AI never reads or retypes your rows.
Two Lose It! specifics are handled deliberately. Its export carries a deleted flag, and rows marked deleted are skipped rather than imported: bringing them back would resurrect food you removed on purpose, and no total on the preview would reveal it. It also writes the literal string “n/a” for cells with no value, which is read as empty rather than as a zero — so a macro you never tracked stays absent instead of being recorded as a real 0 g and dragging your averages down.
Run it as often as you like. Each row carries a content fingerprint, so a repeat import of the same file reports the meals as already logged and adds nothing. And if the dates in your export could be read two ways — 05/06 being May or June — the importer shows its reading against a row from your own file and asks you to confirm it before writing.
How to switch
Works with any MCP client that supports OAuth 2.0 with PKCE. On first connect you create an account with Google or an email and password.
https://your-domain.com/mcp
into the
Remote MCP server URL field and
click Add.
Using ChatGPT or another client instead? The full install guide covers ChatGPT, Cursor, VS Code, Claude Code, and more.
FAQ
No. Lose It! does not offer a Model Context Protocol (MCP) server, so there is no official way to connect it to Claude, ChatGPT, or other AI assistants. Nutrition MCP is a free, open-source alternative built as an MCP server from the ground up, so you can log meals and macros directly inside your AI.
There is no official Lose It! connector for Claude, because Lose It! has no MCP server or public MCP integration. The closest option is Nutrition MCP, a free MCP server: add https://your-domain.com/mcp as a custom connector in Claude, sign in, and start logging by conversation.
Yes — send a photo of your plate and your AI identifies the food and estimates the macros, then logs it after you confirm. In Lose It! photo logging sits behind a paid plan; with Nutrition MCP it's free and works right in the chat.
Yes. The core loop is identical — say what you ate and get your remaining calories and macros back instantly. The difference is you talk to your AI instead of tapping through an app, and there are no ads or upsells on the way.
If you want to track calories, macros — fiber, total sugar, and caffeine included — water, and weight without opening a separate app or searching a food database, yes. Instead of tapping through a database, you describe what you ate in plain language, send a photo, or scan a barcode, and your AI logs it — completely free and open source.
Yes. Ask to import and an importer opens in the chat: you pick the CSV Lose It! exports, it's parsed in your browser rather than read by the AI, and you confirm a preview before anything is written. The date, food, meal, calories, protein, carbs, and fat map themselves, and fiber, total sugar, and caffeine do too when your export carries them. Lose It!'s export is recognised by name — rows flagged as deleted are skipped instead of resurrected, and its “n/a” cells are read as empty rather than as zeros. Re-importing the same file never creates duplicates. In clients without in-chat panels you can paste your export instead.
Not when the importer opens. It parses the CSV in your browser and shows you what will be added before anything is written: how many meals, the calorie total, anything it had to flag, and the rows themselves — a long file lists the first of them plus a count of the rest rather than every line. Only the rows you confirm are sent, and they go as structured data rather than through the AI's reply, so no row can be mistyped or invented in transit. Each row also carries a content fingerprint, so running the same file again reports those meals as already logged instead of duplicating them. If your client can't display in-chat panels, the fallback is to paste the export — the AI does read it on that path, so prefer the importer when you have the choice.
Yes. Nutrition MCP is completely free with no premium tier, ads, or paywalled features — unlike apps that put some features behind a subscription. You only need a Claude or ChatGPT account to connect.
Free and open source — no Lose It! account, no app to open.
Lose It! is a trademark of its respective owner. Nutrition MCP is an independent, open-source project and is not affiliated with, endorsed by, or sponsored by Lose It!. Comparisons reflect publicly available information at the time of writing and may change.