Tracking a Lean Bulk in Claude

Most calorie tracking advice is written for people eating less. Gaining has a different shape, and treating it as a cut with the sign flipped produces one of two bad outcomes: a surplus so small that nothing happens for three months, or one so large that most of what arrives is fat.

The tracking setup is the same as for a cut. The decisions are not.

This is the setup and the loop. It is not a case study. We have not run a bulk on this stack and published the log.

The surplus is smaller than people think

The intuition is that gaining requires eating a lot. For rate of gain, it does not, and overshooting has a real cost.

The rate of muscle gain has a ceiling set by training age and biology, and it is slow, slower for someone who has been lifting for five years than for someone in their first six months. Calories above what that ceiling can use do not become more muscle. They become fat, which you then have to diet off, which costs you time and some of the muscle you just built.

So the commonly recommended approach is a modest surplus, often framed as roughly 10 to 20% above maintenance, with a target rate of gain expressed as a small percentage of bodyweight per month rather than per week. Trained lifters are usually advised toward the lower end of both.

Get maintenance from the tool and add to it:

You: I am 27, male, 74 kg, 179 cm, lifting five times a week, on my feet at work. Maintenance, then a modest lean-bulk target.

One calculate_macro_targets call. As always, that output is a formula estimate rather than a measurement, and the correction comes from watching the scale over weeks. The calculation itself is documented in the TDEE and macro calculator guide.

Protein does not go up much

A common assumption is that a bulk needs dramatically more protein than a cut. Published guidance does not really support that: the range people cite for muscle gain, broadly 1.6 to 2.2 g per kg, overlaps heavily with the range for a deficit, and the upper end is arguably more important when cutting, because that is when muscle is at risk.

What changes in a surplus is that hitting the protein target gets easy. You have calories to spare. So protein stops being the binding constraint, and something else becomes the problem.

The actual constraint is appetite

This is the part nobody warns you about. In a surplus, the difficulty is not discipline, it is volume. A day at 3,200 kcal of chicken, rice, vegetables and fruit is an enormous amount of food, and by week three of eating it you will not want it.

Which flips a rule from the cut playbook. In a deficit, calorie-dense food is a hazard, because it is easy to overeat. In a surplus it is a tool, because it lets you hit a target without being uncomfortably full.

Things that help, and that a catalog-backed planner is good at finding:

Ask for density, not volume. The query is a real one the tools can answer well, because calories and protein per 100 g is exactly what a catalog lookup returns:

You: From my core food list, which items give me the most calories per 100 g, and which give the most protein per calorie? I need to add 500 kcal without adding much volume.

Drink some of it. A blended drink with oats, milk, nut butter and fruit is several hundred calories that takes two minutes. Total it as a recipe, as in recipe macros without a spreadsheet, because a smoothie is the food people most reliably underestimate, which in a surplus is a feature rather than a problem as long as you log it honestly.

Use fat to fill gaps. Oil, nuts, nut butter. At roughly 9 kcal per gram, small weights close large gaps without volume.

Reduce the fibre and raw vegetable volume a little. Not to zero, since fibre matters, but a kilogram of raw salad is a lot of stomach space doing very little work toward the target.

A worked surplus day

The structural problem becomes obvious when you lay one out. Take someone at 74 kg who needs roughly 3,100 kcal and around 145 g of protein.

Built only from lean whole foods, that day is: 300 g of chicken, 400 g of cooked rice, 200 g of vegetables, 250 g of skyr, 150 g of oats, three pieces of fruit, and it still does not get there. It is also about three kilograms of food, most of which has to be eaten in the evening because nobody wants that much at breakfast.

Built with density in mind, the same targets come from: the same protein sources, a smoothie with oats, milk, banana and 30 g of peanut butter, olive oil on the vegetables rather than steaming them, whole milk instead of skimmed, and a handful of nuts. Less volume, same numbers, and it is a day you can repeat on a Tuesday.

The figures above are structure rather than a transcript. Your rows will differ by brand and by which catalog entries resolve. What holds regardless is the shape: two or three density decisions do more for a surplus than any amount of planning precision.

The one thing to watch is that density makes both directions easy. In a cut, an unlogged handful of nuts is a hazard. In a surplus it is a help, right up until the surplus is twice what you intended because three such handfuls never made it into the log.

Training is the other half, and the log cannot see it

Worth stating plainly, because a food log invites the belief that food is the variable.

A surplus without progressive training is weight gain. The calories permit growth; the training is what asks for it. If the scale is moving and your lifts are not, the surplus is going somewhere you did not intend, and no adjustment to the macros fixes a programme that is not progressing.

None of this is visible to the nutrition tools. There is no training log in the eight tools and no way for the server to know whether you trained. If you want the assistant to reason about both together, the training log has to go in your file alongside the food log. A line per session with the main lifts and loads is enough, and it makes the weekly review considerably more useful, because "weight up, lifts up, waist stable" is a conclusion and "weight up" on its own is not.

The weekly loop is slower

Same daily logging, same weekly review, but the decision cadence stretches out.

You: Read the last 28 days. Weekly average weight for each of the four weeks, average calories, average protein, days logged. Then tell me whether to hold or add.

Four weeks, not two. A gain of a few hundred grams a month is smaller than normal weight fluctuation, so a two-week window cannot distinguish it from noise. You need a longer baseline than a cut does, and this is the single most common mistake: people add calories after nine flat days and end up gaining much faster than intended.

Adjust upward in small steps. If four weeks are genuinely flat, add 150 to 200 kcal and give it another four. Doubling the surplus because progress is slow is how a lean bulk becomes a regular one.

Watch the composition of the gain, roughly. The scale cannot distinguish muscle from fat, and neither can a food log. Waist measurement alongside weight is the crude tool everyone uses: weight up and waist stable is the outcome you want; both climbing steadily means the surplus is too big.

Strength is the other signal. If the scale is moving and the lifts are not, the surplus is not being used for what you intended.

Where the log tends to fail in a surplus

Different failure modes from a cut, and mostly the mirror image.

Under-logging becomes over-reporting. In a deficit people quietly under-report. In a surplus the same bias means you think you ate 3,200 and you ate 2,900, so nothing happens and you conclude you are a hard gainer. The four-day strict-weighing diagnostic from the cut post works exactly the same way here, in the opposite direction.

Missed meals do not average out. Skipping a 700 kcal meal on a busy day is a 700 kcal hole. Over a week that is a tenth of the surplus gone. A cut tolerates a skipped meal; a bulk does not.

Liquid calories are now your friend and still need logging. The thing that makes a surplus easy is also the thing easiest to forget to record.

Training has to actually be there. Eating in a surplus without a progressive training stimulus is just gaining weight. Nothing in a food log tells you whether your programme is working.

When to stop

Decide in advance, the same as with a cut. A weight, a duration, or a waist measurement you will not go past. Bulks without an end condition run on until someone notices, and then the cut needed to undo it is longer than the bulk was.

Connect it

Claude Code:

claude mcp add --transport http calorie-api \
  https://calorieapiadmin.com/mcp \
  --header "X-API-Key: YOUR_KEY"

Cursor:

{
  "mcpServers": {
    "calorie-api": {
      "url": "https://calorieapiadmin.com/mcp",
      "headers": { "X-API-Key": "YOUR_KEY" }
    }
  }
}

Browser sign-in for Claude apps is off on this server until OAuth is enabled. The MCP page is the overview, the docs the reference, MCP pricing the plan and 7-day trial.

One practical note on call budget: a surplus day has more food in it, so more entries, so more calls, and a big day can run past 20 tool calls. Still comfortable against 20 a minute and 10,000 a month, but a reason to lean harder on saved core foods and saved recipes so that most entries cost nothing at all.

What we have not measured

We have not run a bulk on this setup and published the log. No rate of gain, no composition estimate, no comparison of logged against strictly weighed intake, no outcome.

What is above is mechanism plus published general guidance on surplus sizing and protein, presented as ranges because that is how the guidance is stated. The claim that calorie density is a tool rather than a hazard in a surplus is reasoning from arithmetic, not a measured finding.

Related

• What the MCP plan includes
• Pricing and the 7-day trial
• Tool reference
• Running a cut in Claude
• Meal planning with Claude and a real nutrition database
• A weekly grocery list built from a calorie target