Factorio Forums

I want to preface this with the fact that I am by no means expert or would consider myself even remotely good at factorio. However I feel somewhat strongly about my experience with Gleba and maybe this feedback will be valuable somehow.

Gleba to me generally has a wonderful puzzle and I am all for spoilage mechanics, but the snag I'm noticing is that it supports only a very specific building style - building it in advance or building it fast. I'm a slow player and I generally either build as I go or have long sits looking at a problem this way and that with scope of the problem being limited. First thing I tried is having something like a bus with heating towers disposing of leftovers at the end, but any problem in between would cause fruit to be stopped processing and seeds to dry up. I feel like some dedicated process that allows to increase seed chance at a worse mash/nutrient output would benefit people like me (if there are any that is) and give more room for experimentation while being secure in knowledge that I have a seed stockpile and don't have to run around exploring to find fruits.

It might be a very niche problem or maybe a problem only for a slow newbie, but it was a problem enough for me to log in and write this, so maybe it's a worthwhile topic for discussion?

The solution I took for that issue is to first build a tiny setup that always processes fruits and mostly burns the non-seed result, so you are left with a slowly piling up of seeds and can just grab some seeds from a chest whenever you need them for whatever reason, be it a mistake or a new build.

I did the same with the pentapod eggs. Have a small slow auto production that voids excess but keeps some fresh eggs around at all times. You can avoid this with a recycler, but you may not have that on Gleba, so this can be helpful.

For the seed surplus, it's important to not let the fruit spoil and always process them beforehand, you can have the mash or the jelly spoil, that's ok, you have extracted the seed first. This is to avoid waste, but to build up an actual surplus, you either need productivity module in the chemplant or use the biochambers which have a productivity bonus without needing module.

You can also limit the agricultural tower with circuit, to prevent harvesting of new fruits if you don't need them, or "limit" it. If you have "trees" they don't spoil, and you can harvest them whenever you need them.

Something I would also like to add is to not overproduce anything on gleba. Under produce. The last machine on the belt should not be running, it's there to eat the overflow. That way, you minimize spoilage.

Another problem I had is, unless im exporting the research, gleba would seize up, making me have to go there to restart it. Instead I built a ship just for gleba, running research to nauvis even when I dont need it, any that spoils I just end up burning there. Lastly I would suggest making the system a loop, so that there's never a backlog at the end causing problems. Thats the only tangible advice I can think of, my gleba is pretty spaghetti/sushi otherwise, but it works
I thought I'd include a picture for some context. Ironically my base is seized but when its running it doesn't stop unless research backs up

println wrote: Sat Jul 04, 2026 8:19 pmI feel like some dedicated process that allows to increase seed chance at a worse mash/nutrient output would benefit people like me (if there are any that is) and give more room for experimentation while being secure in knowledge that I have a seed stockpile and don't have to run around exploring to find fruits.

You can actually make a dedicated process yourself: a small factory whose sole purpose is to generate surplus seeds. So it just harvests trees, processes the yumako/jellynut, replants seeds, then sends the mash/jelly off to be disposed of. A heating tower is simplest, but it could be a production facility of some sort if you can ensure that the disposal belt will never back up. (e.g. use a priority splitter so that if the belt would back up it goes to a heating tower solution)

Here is a very minimal demonstration of the idea: (Note that the tower-to-tower inserter is a blacklist on yumako, not whitelist; it's hard to see in the screenshot) In reality you'd want to maximize productivity bonuses so that you generate surplus faster, and probably want to use more assemblers and planting spots, since the process is somewhat slow.

This is designed to be a separate, isolated factory, to make it that much harder to accidentally introduce a failstate into the design.

But if you can do it reliably enough for your interests, you really should use biochambers because the +50% intrinsic productivity bonus is a big deal. You can even do it on the yumako side while keeping it as an independent factory since you can use mash-to-nutrients to fuel the biochambers.

You might even try using this as a design principle for your main factory: to try and engineer the fruit/jellynut processing to run reliably no matter what is happening further along in the factory.

The best Gleba tip, as someone whose favorite planet it is: Only three items on belts, the rest is inserted directly or transported by bots:
Belt items: Bioflux, Spoilage, Nutrients (in descending order). Try to directly insert nutrients as much as possible.
Always direct insert anything to do with fruits, their lifetime is so short, they should never touch a belt.
Bioflux and spoilage are about the only things that can safely and easily be put on belts.
Raw fruits especially can be handeled best by bots, endproducts too, but belts are fine too if you prefer.

worph wrote: Sat Jul 11, 2026 8:31 pmRaw fruits especially can be handeled best by bots,

I actually prefer trains, but with very different paradigm than the 'usual' usage: small loading/unloading buffers, small inventories, more trains making frequent stops that leave shortly after arrival.

Maybe with a lot of bot speed research I might change my opinion, but before that I feel like properly configured trains do the best at reducing buffer sizes and transit times.

I always wonder how, after just landing on gleba, people manage to build proper tracks to their trees.
Everytime I bootstrapped Gleba it was the easiest and most straight forward solution to just put down 5-10 roboports in each direction and 8 chests.
In my experience there was never enough space or available ressources to justify a complex (elevated!) rail infrastructure.

Plus, once you put these ~30 buildings down and have them powered by rocket fuel there is really no incentive to build trains for a long time.
And it feels so much cheaper to build the robo infrastructure locally or transport it via space, compared to train stuff.

Hurkyl wrote: Sun Jul 05, 2026 2:16 pm

println wrote: Sat Jul 04, 2026 8:19 pmI feel like some dedicated process that allows to increase seed chance at a worse mash/nutrient output would benefit people like me (if there are any that is) and give more room for experimentation while being secure in knowledge that I have a seed stockpile and don't have to run around exploring to find fruits.

You can actually make a dedicated process yourself: a small factory whose sole purpose is to generate surplus seeds. So it just harvests trees, processes the yumako/jellynut, replants seeds, then sends the mash/jelly off to be disposed of.

Seeds don't spoil. So there's no reason to ever do this.

worph wrote: Sat Jul 11, 2026 8:31 pmRaw fruits especially can be handeled best by bots, endproducts too, but belts are fine too if you prefer.

High volume long distance transport is the worst use case for bots, and fruits specifically want low latency and small buffers due to spoilage, so trains are also quite bad. Belts are ideal for raw fruit in my opinion.

nzer wrote: Mon Jul 13, 2026 8:28 pm

Hurkyl wrote: Sun Jul 05, 2026 2:16 pm

println wrote: Sat Jul 04, 2026 8:19 pmI feel like some dedicated process that allows to increase seed chance at a worse mash/nutrient output would benefit people like me (if there are any that is) and give more room for experimentation while being secure in knowledge that I have a seed stockpile and don't have to run around exploring to find fruits.

You can actually make a dedicated process yourself: a small factory whose sole purpose is to generate surplus seeds. So it just harvests trees, processes the yumako/jellynut, replants seeds, then sends the mash/jelly off to be disposed of.

Seeds don't spoil. So there's no reason to ever do this.

I'm not sure what chain of logic leads you to think aobut spoiling seeds, but I quoted the reason to do this -- to have a seed stockpile -- and the primary application is the opening post of this thread -- for people having problems with factory design that "cause fruit to be stopped processing and seeds to dry up".

High volume long distance transport is the worst use case for bots, and fruits specifically want low latency and small buffers due to spoilage, so trains are also quite bad. Belts are ideal for raw fruit in my opinion.

You can use small buffers with trains.

The problem with belts is that it is easy to over- and underproduce. Bots will regulate themselves more easily and autonomously.
When using belts you need to finetune how often you harvest, or other aspects and keep different consumption speeds in mind (ideally). For when you are producing secondary products besides science (plastic, ores, chemicals, etc.).
Now of course one could make a factory that constantly produces a max amount of fruits and always consumes or burns all of them, no matter what.
But that is one of the worst strategies, before you have proper defenses up, meaning: Artillery, tesla nd missile turrets. Which is again much later in the game usually. Because the constant harvesting and processing of excess fruits will create insane pollution clouds and attract constant attacks.
When I use bots and keep my science production low I can essentially guarantee to never be attacked on gleba, whatsoever. Sure this can be one with belts, too. And using tier 4 belts might even give you better freshness, but it's much easier to fail and create more headaches than losing a few % freshness.

Hurkyl wrote: Mon Jul 13, 2026 9:47 pmI'm not sure what chain of logic leads you to think aobut spoiling seeds, but I quoted the reason to do this -- to have a seed stockpile -- and the primary application is the opening post of this thread -- for people having problems with factory design that "cause fruit to be stopped processing and seeds to dry up".

I'm trying to be charitable, I guess, because setting up a subfactory to process fruits and then just throw away the mash/jelly is so horrendously wasteful.

If you don't want your seed production to be dependent on nutrients, you can just do it with assembling machines instead of biochambers inside your real factory.

worph wrote: Tue Jul 14, 2026 1:31 am The problem with belts is that it is easy to over- and underproduce.

Not really. It's trivial to wire your intake belt to the ag towers and set them to harvest when the buffered fruit is below some level. That's more or less the same thing you do with bots to prevent overproduction.

nzer wrote: Fri Jul 17, 2026 12:10 am Not really. It's trivial to wire your intake belt to the ag towers and set them to harvest when the buffered fruit is below some level. That's more or less the same thing you do with bots to prevent overproduction.

It's also much more of a pain to setup belts on Glebas landscape IMHO. But I will admit, it's probably a very similar outcome to wire it like you suggested. Much more energy efficient, too. I might transition my base to belts actually. But personally I still think bots are more confortable at the very beginning, when you don't come to the planet with a finished factory. It's so much less headache to bring and place some roboports, powerpoles and chests. Also feels better to bring a handful each of those item rather than a couple hundret belts and undergrounds and hoping they are enough. But I like bootstrapping all the planets with just a couple bots every time.

Ragarding trains: I still interested to see the rationale or an actual playthrough where someone sets of trains first on Gleba, while bootstrapping it.

nzer wrote: Fri Jul 17, 2026 12:10 am

Hurkyl wrote: Mon Jul 13, 2026 9:47 pmI'm not sure what chain of logic leads you to think aobut spoiling seeds, but I quoted the reason to do this -- to have a seed stockpile -- and the primary application is the opening post of this thread -- for people having problems with factory design that "cause fruit to be stopped processing and seeds to dry up".

I'm trying to be charitable, I guess, because setting up a subfactory to process fruits and then just throw away the mash/jelly is so horrendously wasteful.

If you don't want your seed production to be dependent on nutrients, you can just do it with assembling machines instead of biochambers inside your real factory.

The optimization goal for what I demonstrated is ease of correctness -- the optimization goal is to be maximize the likelihood you have something correct that doesn't have any flaws, subtle edge conditions, or any of the sort of thing that would make things break and run you out of seeds.

If you are not having trouble keeping a seed reserve even in the face of failure, it's not meant for you.

And it was intended to be a fairly minimal proof of concept. If a person is sufficiently confident they can, for example, run the mash/jelly to a more productive factory under normal conditions and still engineer in the failsafe to ensure any problems down that line won't break this unit, then I agree they should do that.

Burn the excess. Always produce seeds
Gleba is infinite and there is no reason to stockpile resources. You unlock burning tower for a reason there. In this situation you will have good freshness