Servitor Craft and Management

Almost every book on servitors tells you how to build one. Very few tell you how many you can afford, and the absence is not an oversight so much as a commercial decision. A manual that says you can run four indefinitely is more satisfying to read and less honest than a manual that says you can run one properly and two adequately and that the third is already costing you something you have not noticed. The reason for the ceiling is simple and unglamorous. A servitor is a standing instruction to your own attention. Maintaining one costs attention continuously, at a low background level, for as long as it exists. This page covers the machinery honestly: where the vocabulary came from, what feeding actually involves, how to count what you are already carrying, how to dismiss one properly, and how to tell a working servitor from a well-maintained fantasy.

A piece of your attention given a job

Strip the framing and a servitor is a behavioural commitment with a symbolic wrapper. You have a task you do repeatedly and do badly, and instead of resolving to do it better each time you consolidate the intention into a single object, charge it once, and then invoke it at the moment of action. The wrapper makes the commitment easier to hold and easier to remember, and it is genuinely useful for that.

This framing matters because it sets expectations correctly. You are not delegating. Nothing is doing the work while you sleep. What changes is that the decision to act has been made once, in advance, in a form you can reach quickly, rather than being remade every morning when you are tired and the task is unpleasant. The entire benefit lives in that pre-commitment, and it is available without any supernatural premise.

Where the vocabulary came from

The load-bearing concept comes from older protective-spirit material: the idea of a designated guardian attached to a person or a place, kept alive by regular attention, and formally released when no longer wanted. Ceremonial literature of the nineteenth and early twentieth centuries picked this up and elaborated it, and Austin Osman Spare developed the version most modern practice uses, framing the entity as a displaced part of the self available for a specific job.

The vocabulary of charging, feeding, and dismissal came along with it, and that vocabulary is well chosen. A charge is a commissioning, a feeding is a maintenance acknowledgement, and a dismissal is a termination with a record. Anyone building a servitor from older sources will recognise the structure immediately, because the older sources had the same shape. The difference is that the older sources tied it to a spiritual patron, and modern practice ties it to attention.

What feeding actually is

Feeding is the step most likely to be misunderstood, because it is described in older sources as an offering and gets read as a sacrifice. It is not. In the modern form, feeding is the deliberate visualisation of the task being completed, done at the moment you complete it. You are rehearsing the done state, which is the psychologically meaningful part, and the token is the visualisation.

The useful consequence is that feeding and doing are the same act. You do not feed and then work; you feed by working. This is a considerably better design than a system requiring separate offerings, and it means the servitor's maintenance cost is close to zero for a task you were going to do anyway. A servitor attached to a task you do not perform has nothing to feed on, which is a good diagnostic.

The ceiling nobody puts in a book

The folklore says a practitioner may maintain a fixed large number of servants. The practitioners who report actual long-term experience report a ceiling far below that, and the reason is the one above: each active servitor is a background process. You are not thinking about it, but you are maintaining the conditions for thinking about it, and the cost does not appear as concentration. It appears as a slight background restlessness and a growing reluctance to start anything new.

The practical figure most people converge on without being told is one to three. One is comfortable. Three requires you to be genuinely organised. Beyond that, the returns collapse, because each servitor is receiving less maintenance attention than it needs to stay coherent, and an incoherent servitor is worse than no servitor, since it adds cost without adding function.

Counting what you are already carrying

Before building anything, list what you have already committed to. Not servitors, but standing commitments, ongoing projects, and people who depend on you. Then mark which of those you are consistently neglecting. The neglected items are the candidates, because that is where the gap between intention and behaviour is widest, and that is exactly what a servitor closes.

This exercise usually produces a shorter list than people expect and a better one. It also tends to surface a commitment you cannot or will not drop, which is a genuinely useful discovery independent of any occult work. If the list comes back with twelve items and no clear priority, the problem is not servitor management. It is that you have not decided which things matter, and building a servitor will not decide it for you.

The first servitor, built in one evening

The build takes about twenty minutes. Choose one task that is repeated, specific, and currently done badly. Write the description in a single imperative sentence with a countable completion condition, so that done is verifiable rather than felt. Reduce that sentence to letters, cut the letters into a mark, and write the office on the back of the page in the same words you will use later.

That is the whole build. Nothing in the method requires an elaborate design, and elaborate designs are the most common reason first servitors fail. The failure mode is spending two hours making a beautiful mark and then never using it, because the making felt like progress and the using does not. A plain mark that you invoke daily is worth more than a beautiful mark that sits in a notebook.

Charging it and the first feeding

The charge is ninety seconds of holding the mark while reciting the office sentence once, and it is done cold, in an ordinary room, without ceremony. The state you want is not belief or intensity. It is a clear connection between the mark and the sentence, so that seeing the mark later recalls the job. If you cannot say what the mark is for within a second of looking at it, the charge did not work.

The first feeding happens the next time you do the task, and it happens by doing the task. At the end, before you leave the desk, spend ten seconds visualising the completed version. That is the feed. Do this for a week and the association begins to set, and the mark becomes genuinely faster to reach than the original description was.

The dismissal, and doing it properly

Dismissal is the most under-taught part of the craft and the one that determines whether your long-term practice is sustainable. The method is short: name the office, state that the commission ends, name the date, and then destroy or bury the mark. The destruction is not symbolic, it is housekeeping, because a mark left in a drawer is an active commitment you have not actually ended.

The discipline is to write the dismissal down, with the date and the reason, in the same notebook where the build is recorded. This serves two purposes. It gives you a record you can audit later, and it imposes a small delay, which is exactly what you want before terminating something you built during a bad week. Most premature dismissals happen on a difficult day and most are regretted.

Signs it is working, and signs you are

The real sign is boring: the task gets done more often. That is the whole outcome, and if it is happening, the construct is doing its job regardless of what you believe about it. Secondary signs are that you stop needing to re-read the office before acting, that starting the task takes less effort than it did, and that you can hold the mark's meaning across a gap of days.

The signs that it is not working are equally identifiable and usually misread. If you feel a persistent low-grade obligation that makes new projects harder, the servitor is consuming more attention than the task returns. If you check the mark repeatedly without acting, the association has not set and the maintenance is not running. If the task gets done but your sense of having done something occult has faded, the servitor is working and your attachment to the framing has not, which is a normal and harmless stage.

Where servitor practice goes wrong

Three failures dominate. The first is multiplication, where the practitioner builds several because building is enjoyable, and ends up with more standing commitments than they can maintain. The second is abstraction, where the office is written as a general improvement to their life rather than a specific task, so there is no completion condition and therefore nothing to feed. The third is inflation, in the older sense, where the practitioner comes to identify with being a person who commands spirits and gets less from the actual work.

The third is the one to watch for, because it is the one that does not announce itself. It feels like progress. A practitioner deep in it will describe increasingly elaborate servitor systems, will be reluctant to do the boring physical work the servitors were built to handle, and will be genuinely puzzled when the results dry up. The correction is unglamorous: stop building, do the task by hand for a week, and see whether the task actually got harder.

Keeping the count down

The management practice that works is a quarterly review, and it takes twenty minutes. List the active servitors with their build dates. For each, mark whether the underlying task was completed in the last month. Any servitor whose task was not completed once is either badly matched or unnecessary, and both cases call for dismissal. Any servitor whose task was completed without the mark being used has outgrown its own tooling and can be released.

This review will occasionally produce a list of zero, and that is a legitimate outcome. Some quarters the right answer is that you have no active servitors because the commitments in your life changed. Holding an empty list is not a failure of practice. It is an accurate description of the current state.

The retirement you should schedule

Every servitor should have an expiry date written on the build page before it is charged. Three months is a reasonable default, extendable once, and the date should be set at build time rather than decided later. Setting it in advance does two useful things: it bounds the maintenance cost, and it removes the decision from the moment when you are most invested in keeping it.

The habit this builds is the actual point. A craft in which everything you build is temporary, reviewed, and released produces a person who finishes things. A craft in which servitors accumulate produces a person with a drawer of unfinished marks. The retirement date is the difference between the two, and it costs nothing to adopt.

Frequently Asked Questions

How many servitors can I maintain?

Most people who keep long-term records converge on one to three, well below what the older literature implies. Each active servitor costs a small amount of continuous attention, and that cost shows up as background restlessness rather than as difficulty concentrating. Above three, the returns drop sharply because each one is receiving less maintenance than it needs to stay coherent.

What is the difference between a servitor and a tulpa?

A servitor is built for a task. It is commissioned, charged, fed by completing that task, and dismissed with a record. A tulpa comes from a different tradition and is built through sustained visualisation over weeks, without a job attached, and it is a person-like construct rather than a behavioural commitment. The servitor is a commitment device with a symbolic wrapper; the tulpa is a practice of disciplined mental imagery.

What happens if I neglect one?

It fades, which is normally harmless. The cost of a neglected servitor is the maintenance attention you keep spending on the idea of it, without getting the function. A neglected servitor should be formally dismissed rather than left in a drawer, because a mark in a drawer is a standing commitment that never got closed, and those accumulate faster than anyone expects.

Does the mark need to be beautiful or intricate?

No. The mark is a retrieval key, and what it retrieves is the office sentence, not the visual. A plain mark that you can produce in ten seconds and recognise at a glance will outperform an elaborate one you have to hunt for, because the invocation has to be fast enough that you actually do it. Complexity in the mark is a comfort, not a mechanism.

Can a servitor act on someone else?

It should not, and the reason is not the metaphysics. A servitor built to influence another person is a standing instruction to your own attention aimed at a target you do not control, which means every time it does not work you will be tempted to increase the intensity rather than examine the premise. Keep the servitor pointed at your own behaviour. The indirect effects on other people follow from your changed behaviour, not from the construct.