How EatMove works

A nutrition and training log with a real dashboard, a real database, and a door your own AI assistant can walk through.

What it tracks

Three things, deliberately: what you eat (meals with calories and macros — protein, carbs, fat, fibre, sugar), what you do (a weekly training template, programs with sessions, one-off activities, and how much you actually burned), and how your body responds (weight, body composition, circumferences, and benchmark efforts like a 5 km run).

No food database to scroll, no barcode scanner, no streaks. If you want those, there are excellent apps for them. EatMove is for people who'd otherwise be keeping this in a spreadsheet — and want the spreadsheet to talk back.

The day model

Every day is a small ledger of events — a breakfast, a lunch, an evening session, your resting burn. Events come from you tapping them in, from your AI logging them on your behalf, or (later) from a wearable. Each event remembers where it came from, so nothing is ever double-counted and everything can be corrected.

Events roll up into the day's totals: calories in, calories out, and your macros against target. That rollup is what the Today view shows and what your AI reads when you ask "how am I doing?".

Targets that change with the day

Most trackers give you one flat number for every day. Real training doesn't work like that. EatMove knows whether today is a rest, resistance, cardio or combined day — from your weekly template — and your macro targets follow. A push day gets more carbs; a rest day doesn't. You set the numbers once per day type and stop doing mental arithmetic.

Connecting your AI

EatMove speaks MCP (Model Context Protocol), the open standard AI clients use to call external tools. That means your assistant can read and write your log directly — securely, with your permission, and without you pasting anything into a chat.

Claude (web or desktop)

1. Sign in to EatMove with Google.
2. In Claude: Settings → Connectors → Add custom connector.
3. Paste https://eatmove.fit/mcp and save.
4. Claude opens an EatMove screen asking you to approve access to your log. Approve once.
5. Start a chat and talk: "had porridge with berries", "what's my protein at?", "what's scheduled this week?".

Claude Code

claude mcp add --transport http eatmove https://eatmove.fit/mcp

Other MCP clients

Any client that supports remote MCP servers with OAuth works the same way — point it at https://eatmove.fit/mcp and approve the access screen.

What your AI can do

Once connected, your assistant gets a fixed set of tools — nothing more. Roughly:

• log_meal
• log_exercise_set
• log_ad_hoc_activity
• log_measurement
• log_benchmark
• get_recent_energy
• get_calendar
• get_activity_log
• get_profile
• get_template
• update_template
• create_program
• add_program_sessions
• mark_activity_done
• mark_holiday
• suggest_benchmarks
• generate_plan

Every tool acts only on your account, only with the scopes you approved, and every write is attributed to the agent that made it so you can see — and undo — what it did.

Your data and your control

• It's a real database. Your full history, not a rolling window. Export is on the roadmap so you can take it anywhere.
• Access is yours to revoke. Every connected app is listed under your profile; one click disconnects it and invalidates its tokens.
• Standard, scoped OAuth. Connectors get short-lived tokens tied to specific permissions (read metrics, write energy, and so on). No API keys to manage or leak.
• No ads, no selling data. You're the customer, not the product.

Bugs, ideas, questions

Email info@eatmove.fit. Bug reports and feature requests are genuinely welcome — a one-person project runs on them.