#Devlog 9

I continued working on rebuilding my project

Recreated the creator agent

I successfully recreated a creator agent and implimented the slides and website generation. Although it is basic, it is a WIP and is a strong foundation.

Rebuilt the TUI to look as close as possible to the original

After the reset, I had to rebuild the TUI, and I decided to just make it look as similar as possible to its original version.

What’s next?

• Lots of polishing
• Extending on the bases of all the agents
• More!

Devlog 8

Went nuclear and did a git reset because smth was killing my code. I still don’t know what, but it is what it is. At least now I’ll be able to fix it and get back to my old stage. Im working on fixing it.

Btw this guy is gonna be the death of me

Devlog 7

hey guys.

seventh devlog. devlog six gave the swarm a memory - metrics, cache stats, handoff index, overdue tasks.

TUI parsing

tui/runner.py parses agent markers now with parse_agent_marker, soprogress is real instead of spinner theater. short_status_line rendersprogress like 2 of 6 with percent plus writer in the footer, and acancel sentinel lets Ctrl+C set a flag file the worker checks insteadof killing mid-write. no more half decks on cancel.

the web app remembers and shares

app.py appends every job to job history json (last fifty kept),so refresh does not wipe your day. zip_outputs bundles pptx, docx,html, xlsx, pdf, and md into one download, and safe_filenamesanitizes anything weird before it hits the header. small endpoints,big usability win.

the guards got pickier

agents/pipeline_guards.py added three focused gates. readabilityfails on filler density and paragraph-bullets, heading structurerequires at least three real sections, evidence coverage requires atleast two distinct http sources. they run cheap, explain themselvesin one line, and catch the exact slop the critic used to wavethrough.

the creator writes decently now

agents/creator_quality.py grew humanize_text_v2 - in order becomes to, due to the fact that becomes because, unlock the power of becomes use. score_section_visual picks timeline, comparison,stats, bento, or text from actual cues instead of forcing one layout. tighten_bullets clips at word boundaries with an ellipsis so slides never overflow their boxes.

the orchestrator cleans up after itself

agents/orchestrator.py quarantines suspect artefacts underquarantine folder instead of clobbering good output, exposesprogress_snapshot for both TUI and web UI, and should_resumedetects a checkpoint worth resuming. run_with_progress wraps nodeexecution with an optional callback so both frontends share oneprogress shape.

tests lock it in

tests/test_swarm_hardening.py covers the whole pass: cost math,empty summaries, filler rejection, heading minimums, url counting,cache stats, missing-handoff detection, marker parsing, and the humanizer scrub. nine tests, all deterministic, no network, no LLM.run them with pytest tests/test_swarm_hardening.py -v.

where it stands

open a fresh deck, site, doc, sheet, and pdf from the same draft -they share a brand, titles are short, bullets fit, images match the subject, motion is subtle, filler is gone. open the session folder -plan, research, analysis, draft, critique, creator, plus index.json.open the metrics - elapsed, hit rate, guard wins. the swarm finally feels like a tool you can run all day and trust.

This is a website it created for a business

Devlog #6

I improved this project by a little bit

Slideshows

Slideshows now have backgrounds, images, graphics, fonts, designs, and nicer templates

websites

Websites now have animations, images, sections, and are customized and tailored to the research task!

documents

Docs now have colors to organize thoughts and sections, and as well as images and visuals every so often

What’s next?

Drastic imrpovement, and new output type soon

Devlog 5

hola

fifth devlog. the last two got the decks real structure and stopped a bunch of problems with like backgrounds and other stuff.

so this one is all about the creator. planner, researcher, writer can do their jobs fine. However, if the creator fumbles the last mile, nobody cares how good the draft was.

the creator improved

agents/creator.py was the biggest culprit. _clean_content_for_slides was stripping file paths and a handful of phrases, but “delve”, “tapestry”, “game-changer”, “seamless”, “robust”, all that stuff was sailing straight onto slides.it strips all of that now. expanded filler list, whitespace collapse, drops those **** divider lines that used to become phantom slides, truncates sanely instead of cutting mid-word.also added three little helpers i wish i’d had earlier. _premium_truncate cuts at a word boundary and adds the ellipsis, _humanize_heading cleans the [layout:...] tags and keeps titles under 60 chars with proper casing, _score_section_visual actually looks at the section — timeline words + 3 bullets = timeline, versus/comparison + 4 bullets = comparison, stats + numbers = stat grid, 5+ bullets = bento. no more forcing every section into the same text wall

pptx is finally actually semi-decent with animations and stuff

this one bugged me for a while. agents/pptx_builder.py had _add_shape_animation and it was literally a no-op. just return. so we’d promise motion and ship static decks. now it wires real entrance animations into the timing tree. The animations are appear, fade, fly, grow, wipe, with delay support. it’s still subtle, staggered, not the 2007 bounce-everything-in kind. charts ease in, bullets cascade, hero elements land in order. open it in PowerPoint and it actually feels presented, not pasted.

the critic got pickier

agents/critic.py would pass stuff with “showcasing” and “unlock” all over it as long as the structure was right. added _critic_visual_score, which checks min 4 sections for rhythm, flags filler hits, flags bullets over 110 chars, needs 82+ to pass. it’s caught three decks this week i would’ve shipped before.agents/polisher.py got the same strictness on the html side — enforces hierarchy, fixes those 10px body styles that slipped through, strips leftover ai-generated / placeholder / lorem classes.agents/visual_builder.py got cinematic helpers, so it has proper 21:9 captions, glass CSS token, bento span logic so the first cell gets the weight.

where it stands

opened five fresh outputs side by side yesterday — deck, site, doc, sheet, pdf from the same draft. for the first time they felt like the same brand. titles are short, bullets fit, images match the subject, motion is there but doesn’t shout, and i didn’t spot a single “delve” or “tapestry”.

Devlog 4

Drastically improved the slideshow generator and website generator

Not much here, just took me 6 HOURS TO FIGURE OUT WHY THEY SUCKED

The agent didn’t have the capabilities to do a background, so it was always white. Also, it was stickign to very basic styles and templates. I made them so much better (still a WIP)

Last 3 hours were spent actually working on and off

Only the last 3 hours were actually spent doing real work, so I lowk am cooked.

Here’s some fried rice and smoothie I made! The tanghulu was a failure but at least it tasted good :P

Devlog 3

hey guys. Third devlog. Devlog two shipped fifty-two templates and an art director,and the html pages finally looked somewhat decent. However, they still sucked. The slides on the first iteration had almost no images, no background, nothing. It sucked until I asked it a bunch of times to revise it.

the forensics

before writing a line of new render code i unzipped both decks and diffed the xml. Agentswarm: 15 identical stat slides, 1 picture total. the reference:13 slides, 12 large images, and a real story arc - full-bleed cover,overview dashboard, content slides with big image panels and labeled morning/afternoon/evening blocks, an insider-tip callout, a section divider. the gap wasn’t the fonts; it was structure and imagery. and since our renderer preferred keyless stock chains and fell back to random picsumgarbage when they failed, “the images are buns” was literally true.i also found the deck was leaking its own internals onto the slides:template names, character counts, a signature footer. nothing says”designed by an ai who gave up” like “generated with template vogue”.

the rebuild

scripts/pptx_template_renderer.py got rewritten around one rule: atemplate is guidance, not law. it still picks a palette, typeface pairing,and hero flavor from the spec, but every layout is chosen from the actualcontent. the archetypes are the workhorses now:- full-bleed cover with scrim, eyebrow kicker, giant title, meta strip- agenda slide with real numbered sections and content-derived stat cards- feature slides: big image panel, up to three labeled blocks (the  context / the detail / the takeaway), a key-takeaway callout- timeline, comparison, quote, and gallery variants for sections that  call for them - capped so a deck never spams the same layout- section divider, sources, and a closing slideevery slide gets header chrome with a section badge and page numbers,plus speaker notes. the image pipeline now tries the ai generator firstwith per-section prompts (verified live - one image every ten seconds),then the stock chain, then a styled gradient. the red badge and randompicsum are gone.

where it stands

still deterministic, still no llm in the loop, still pyright clean. the bronze ox is gone because the pipeline finally knows the difference between a farm animal and a model name.

still ahead: teaching the website renderer to generate its own images when the cache is cold, and pushing the same director thinking down into the pdf path.

This is a one-shot result - it kind of sucks

Devlog 2

hey guys.

second devlog. the first was about keeping the pipeline alive underpressure; this one is about making it actually good. the goal was to stop thecreator from reaching for the same handful of preset looks, and to make theresearcher think before it searches. along the way a few hundred errors inthe problems panel just magically fricking appeared and took me way too long to fricking solve.

A bunch of new clean templates

The previous set of templates that the creator agent was given were complete dodo, like worse than the gta 6 leaks by cyberpunk and thick of it. the old template system was a few hardcoded themes and a color swap. Now, there are fifty-two distinct templates across eight categories - editorial, corporate, tech, creative, minimal, data,storytelling, luxury.

Each one is a complete design dna, not a recolored clone: its own palette, typography pairing, grid, signature element(oversized type, noise grain, gradient orbs, marble texture…), and section composition rules. both the html and pptx renderers consume the same specs,so a website and its matching deck look like siblings.

New agent - art director

Templates alone just give you more wallpaper, so the creator got an art director (agents/art_director.py). it maps the task and industry to a visual direction - palette mood, typography voice, imagery style, layout personality - across ten moods from editorial to brutalist to nightlife. a coffee shop brief lands on faded-sun vintage with grainy texture; a saas pitch gets cool slate and a bento grid. no llm, always available, and its direction folds into template selection so outputs stop looking like one fixed style.

Added visuals to the slides (still need a lot of work)

the pptx renderer’s primary path is now the spec-driven template renderer:one dominant focal point per slide - image plus caption, stat, quote,timeline, or comparison - short visible text, and human speaker notes onevery slide. something visual every slide, balanced against the words.

quality passes before anything renders

markdown now runs through deterministic passes in agents/creator_quality.pybefore the creator touches it. humanize_text scrubs the telltale ai-isms(moreover, leverage, delve, seamless…) so the voice reads human.narrative_arc_md makes sure sources land at the end instead of opening thepiece. annotate_visual_hints finds sections that would render as unbrokenprose and injects a takeaway bullet from the section’s own first goodsentence. and the new quality_gate self-reports coverage issues onproduced outputs.

the researcher plans and verifies now

research used to be: search, scrape, write whatever came back. nowagents/research_plan.py turns the task plus the planner handoff into subtopics, key questions, and evidence targets that anchor every search round - targeted instead of spray-and-pray, with deterministic fallbacks so fast mode still works with zero llm calls. and agents/evidence.py extracts atomic claims from scraped sources, then verifies them across sources:corroborated, contradicted, or unsupported. the synthesis prompt now refuses to present a number as fact unless the evidence view backs it, and the researcher wires its handoff text back into the prompts it actually sends -which was fetch-and-forget before.

This is a slideshow it made after a lot of tweaking (like 20+ prompts); I want it to be able to do such a thing in 1 prompt so still a lot of work to be done.

Devlog 1

hey guys.

foundatin for the project

I built a simple foundation for the project with a nice TUI and simple functoinality (base version for each agent).

Researcher agent

The researcher agent so far has a base firecrawl api, simple web search, and does nothing else so far

Analyst agent

The analyst agent’s only job is to analyze and summarize all key info found by the researcher

Writer agent

The writer agent’s job is to write prompts for the creator agent including all relevant information from the research

Creator agent

The creator’s only job is to create slideshows and other outputs based on given info

Critic agent

The critic’s only job is to criticize the output quality of the creator agent

These are all only rough implimentations of them, more is yet to come

Devlog 28 - The Egg Becomes an Altar

hey guys.

last devlog for this project

The egg now hatches

I made the egg hatch into an end golem after 1 minute - a very op, hard to defeat mob. However, defeating it gives 3-5 additional permanent hearts (lose it upon death).
Also, when you kill the dragon now, a chest spawns where it died, with dragon wings (gives iron-level protection, slowness 1 while worn, functions like an elytra).

what that means for the endgame

so the progression reads clean now. beat the dragon, take the egg, set it up,and call up the colossus that guards everything past the first kill. the dragon fight is the warm-up act for something that doesn’t even spawn until you put the trophy to work.

that’s it. that’s the end of the end. thanks for following along - go fight the dragon for real, then come back and raise one of your own.

Devlog 26

hey guys.

I turned the dragon fight into a bunch of small fixes that all made the fights read better.nothing shiny, just the stuff that was quietly wrong.

stuff that was supposed to do stuff but didnt lol

the dragon fight had announcements. not the kind players see - the kind thatonly exist in code. the enrage thresholds (“The Dragon Remembers” at 60%,35%, 15%) and the hoard reveal were computed, and then… dropped. a soundplayed, a title was built, and nobody ever saw the title. the function wasliterally called announce and it announced nothing.

same story in the special attacks. the imports for title packets were sittingthere, imported and never used. so now the scripted set-pieces finally namethemselves:

• Burrow - “The Dragon Descends” before the ground starts shifting
• Seize - “The Dragon Takes You” right before it swoops to carry you aloft
• Rifts - “Reality Tears” as the gravity rifts start opening around you

i’m genuinely annoyed this was broken for as long as it was. the drama of thedragon fight was all there in the code - the escalation, the set-pieces, the named thresholds - and none of it reached the player. now it does.

the sentinel’s grab is escapable now

the crown sentinel’s grab-and-hurl from a few devlogs ago was a one-way trip.if it got its hands on you, you watched the animation and then got launched.no counterplay, which felt wrong for a fight that’s supposed to be about reading the boss.

so it got the same rule the warden and golem already had: deal enough damage mid-hold and it drops you early. 18 damage is the breaking point - enough that you can’t just tickle it, but enough that a player who’s actually fighting back gets out before the throw. it also shields against the worst feeling case, where the hold just deletes you with nothing to do about it.

and while I was in there: the goal list had two goals at priority 4, whichmeant the sentinel’s behavior was occasionally picking the wrong thing to doat the same weight. renumbered so the hierarchy actually means something.

next up

the dragon’s hoard, the sentinel’s other moves, and the stronghold is still missing its advancements - the burrow is found, but it doesn’t have a single achievement attached to finding it. that’s next.

the fights are starting to feel like they have a conversation with you now.the dragon tells you what it’s about to do, and you get to do something about it.

That is the top view of the crown sentinel btw, it looks pretty dope imo

Devlog 25

Wrapping things up

Hi guys, this was mostly wrapping up the project. I have had a lot of fun working on the project, but I want to ship soon, maybe at like 200 hours. I hope everyone has enjoyed watching my progress on this, as I thoroughly enjoyed working on it myself. I hope everyone gets a chance to try it out and give some feedback.

Models

I touched up some 3d models and improved them a bit, making them look a bit cleaner. I touched the void stalker’s model, and the dragon’s model primarily

Loot tables

I customized the loot tables for a lot of the custom or modified structures like the stronghold to be a lot nicer and more fitting towards the endesium mod. They are a lot stronger, but not too OP I’d say.

Nothing much else done

I know I spent 10 hours coding, but I didn’t get much done because of how long I spent wondering what to do lol.

Devlog 24 - Making the Stronghold Even Better

hey guys.

DEVLOG_23 was about replacing the vanilla stronghold - the layout, the rooms, the portal cathedral. and it shipped, and it worked, and then i walked back through it and something was off. the shape was right but the rooms read empty. a big citadel of bare stone is just a bigger version of the problem i was trying to kill. so yesterday/today was the boring pass nobody gets excited about in a change log: making rooms feel lived in.

the noise

every room had the bones but none of the clutter. so i went room by room and dressed them like actual inhabitants left mess behind:

• wall shelf rows with candles on top, sconce rows down the hall, two-tone  floor tiles instead of one flat slab
• clustered corners - reading nooks, work benches, storage piles - instead of
• hanging lanterns and chandeliers where the ceiling actually had headroom  for them
• crypts and catacombs got coffin rows, skull niches, soul lanterns
• barracks got lockers, footlockers, armor stands
• treasury got actual stacks of gold, not a promise of gold
• gardens finally read as gardens - flower meadows instead of a dirt square

and corridors stopped being dead space: candle runs, clutter, tiled edges.it’s not glamorous but it’s the difference between a structure and a place.

three new district wings

once the rooms felt alive, the citadel wanted more than the hub spine. so it grew three themed districts, and i love how they ground the place:

• Forge - a contained lava pit ringed in polished deepslate, anvils,  furnaces, coal stores. the stronghold had no industry before; now it reads  like it smiths something down there.
• Dining Hall - a proper mess: table rows, benches, chandeliers. a  fortress this size would have somewhere to eat.
• Prison - iron-bar cells, chains, a witch behind the barracks. every  good stronghold needs a brig.

plus a monumental columned entrance to the Great Descent Hall, an amethyst landing dais, cross-bridges over the Star well, a grand staircase dropping from the cathedral gallery, and a carpet runner down every corridor so you always know you’re on the path.

Also started playing hades :P

Devlog 23 - The Stronghold Update

hey guys.for a while now the Endesium identity stopped at the end. everything past theportal was ours - the island, the dragon, the resonance, the gear - but the road to it was still vanilla’s stronghold. a cramped maze of hallways and libraries copy-pasted a hundred times, holding the exact same end portal the game shipped with in 2011. it never matched. the mod’s big gate got the smallest room in the game.that’s done now. the vanilla stronghold is gone and the Endesium stronghold replaced it: a buried citadel that previews the end without spoiling it. this one took a while, so let me walk the whole thing.

the layout: a citadel

vanilla strongholds are random. rooms are scattered, corridors dead-end,libraries repeat, and the portal is wherever the RNG put it. the Endesium stronghold is authored: a fixed, connected, three-times-bigger floor plant that always works.

• a Great Descent Hall buries you - a switchback staircase that drops  from the surface as much as 90 blocks down into the stone, red runner  carpet down the aisle, end rods lighting the way, two vindicator sentinels  waiting at the top.

• a sealed Arrival Vault is what the eye of ender actually points at: a  quiet chamber with a radial purple sigil on the floor and exactly one  grounded door - the route onward, always.

- the Hub is the spine: an atrium with a carpeted floor, a central  resonance dais (amethyst block, end rod, the works), purpur pillars at the  cardinal axes, bookshelves and barrels lining every wall, bats circling the  ceiling, and a real staircase up to the gallery instead of a visual  platform.

• junctions are readable - four grounded route markers plus a carpet cross -  so you never need signs to find your way.

the rooms

sixteen distinct room kinds, every one hand-dressed:- Library - bookshelves three shelves high around the walls, two  lecterns, a chiseled-bookshelf reading dais, blue carpet, bats.- Treasury - iron-bar vault bars with a single gap, chests and barrels  in rows, gold, candles, and a captain guarding the middle.- Guard Post - weapon racks (fence posts) with candles, a phantom  spawner in the center, two vindicators on the door.- Sanctum - end stone brick floor, purpur pillars, crying obsidian and a  chorus plant on a pedestal, purple carpet, an enderman spawner, a live  enderman pacing the corner.- Barracks - four red beds with proper head/foot parts, smoker, furnace,  crafting table, three guards on duty.- Scriptorium - lecterns with candles, bookshelves and chiseled shelves,  a cauldron of ink, a bat haunting the rafters.- Arboretum - a glass-ceilinged garden: dirt floor, flowering azaleas,  ferns, moss carpet, a lily pad pond.- Observatory - concentric purpur ring inlays, an amethyst  telescope-core, end-rod instruments along the walls.- Resonance Engine - amethyst conduit pillars with end-rod emitters and a  crying obsidian reactor - the end’s power plant, studied safely from  outside the portal.- Bastion - deepslate pillars, an iron-bar partition, weapon racks, a  captain and two guards. military.- Conservatory - chorus plants and chorus flowers growing on end stone,  flower pots and moss, an enderman in residence.- Reliquary - concentric obsidian ritual rings, an ender chest at the  heart, candles, a captain standing watch.- Portal Cathedral

what’s next

the stronghold finally reads like the end’s doorstep. next up is the rest ofthe ten-hour pass - every file retyped, every line deliberate - and afterthat, the bosses and dragon still want their second pass of polish. but thegate is fixed. you’ll never walk vanilla’s cramped hallways to the end again.

Devlog 22 - End Fight Update (again)hey guys.this one’s a two-day entry. yesterday was the sky getting alive and the gear lines finally clicking together into actual fighting systems. today was the dragon. the end fight has been building layers for a while - phases, enrage,scripted moves - but it still had one big problem: the first dragon barely used any of it. that’s fixed now. the first kill is the fight now. therespawned one is the nightmare.

yesterday - 8/26the sky got three new mobssky jelly, galefin, and deep lurker went in (that one got its own devlog - 21).the sky half of AmbientSkyManager fills the air with biome motes when youfly high, and the deep half keeps the under-island dark honest: lurkersspawned into the black pockets below the heightmap, capped so the deep end isominous without ever becoming a wall.the sets became systemsthe luminous, ash, and null sets each got a cross-set layer inEndgearSetBonuses - wear the full set and hold that line’s tool andarmor + weapon merge into one fighting identity:

luminous + prism grip - speed II and haste II while channelling; every hit blinds with a flash (3 true damage + glowing) and sparks light shards.ash + ember core - fire resistance while channelling; hits stoke the burn for 2 extra seconds and throw weight behind the strike.null + erased edge - strength I and haste I; every hit shreds 4 armor durability off the target and drags them into the blade.
and the void set finally got its signature: hover. full void set, sneak inthe air, and you stop falling - a gentle drift at -0.12 blocks/tick instead ofa tumble, with void motes trailing under you. it pairs with the sword’ssingularity for total sky control, and it’s the boots’ tooltip promise kept.

today - 8/27the dragon actually fights nowthefour-phase fight controller - impact dives, shockwaves, screeches, gales,storms, collapses, the final roar, catastrophe - now runs on the firstdragon too. before, the first kill was vanilla strafing plus harassment, andall the authored grammar was locked behind the respawned dragon. that gate isgone. the first dragon opens with the full attack schedule, escalates throughall four phases, and the transformation still makes things worse: 1.6x scaleand the resonance-exclusive moves are still the awakened dragon’s alone.the pillars became cooler with new mechanicscrystals were just a heal before. now the aegis is a shield:- every surviving pillar grants the dragon 10% damage reduction (capped at 60%), so clearing the pillars is a real first phase, not an optional chore.- destroying a crystal detonates at its pillar with a roar.- the last crystal falling breaks the aegis: the dragon is staggered for five seconds and takes +40% damage, a wisp wave pours out, and the next scripted set-piece fires immediately. that’s your payoff window.

set-pieces are guaranteed nowscripted specials used to roll a dice every tick. now they’re scheduled:one opens the fight about thirty seconds in, every enrage escalationforces the next one within a second, and a broken aegis queues one on thespot. the random cadence only fills the gaps. you will see these moves, everyfight, on purpose.

two new attacksOblivion Charge (enrage 2+, set-piece) - the dragon climbs high, paints a straight lane through the nearest player with end-rod markers, then dives it twice, tearing through anyone on the line and leaving void-fire wakes in its path. get off the lane.- Resonance Collapse (transformed dragon only) - drags every player toward the dragon for three seconds with converging rings, then detonates at its position. gain distance and find cover; standing still loses.

Devlog 21

hey guys.the End’s been empty for a while. the ecology mobs live on the islands, thebosses live under them, but the sky itself was a void. this one fixes that,and it brings back something else that was missing: three new mobs

three new mobsSky Jelly - ambient, drifts with the wind. a soft bell of a thing withfour trailing tentacles, pulsing gently and leaving end-rod motes behind it.it bobs upward every now and then like it’s breathing. 8 hp, doesn’t careabout you. just ambient.Galefin - the fastest thing in the End. a sleek sky-fish that cutsthrough the air at 0.75 flying speed and avoids you entirely - the momentyou get within 12 blocks it’s gone. 4 hp. you see them flicker past in thecorner of your eye more than you ever actually look at one.Deep Lurker - this is the mean one. it lives below the islands, in thedark air pockets under the heightmap, and it waits. 22 hp, leaps at you,does real damage. when one spawns you get a portal-particle burst as awarning, and then you have maybe a second to look down.all three are geckolib animated - jelly pulses, galefin swims, lurker idleswith a slow head sway - and all three have spawn eggs and their own loottables.

the sky and the deep are managed now

AmbientSkyManager runs two server-side systems:sky half - if you’re flying more than 12 blocks above the surface, theair around you fills with biome motes (highland wind, lumen motes, endwastes dust). it makes flying at height feel alive instead of empty.deep half - a managed lurker spawner. every 5 seconds it looks for darkpockets below the islands: solid floor, air to spawn in, brightness under 7,at least 34 blocks of island above you. spawns are capped at 6 lurkerswithin 96 blocks per player, so the deep end gets ominous without everbecoming a wall of them.

what’s nextthe deep end is less empty now, but it’s still just lurking. the design docskeep pointing at something bigger under the islands - the Umbral Reach - andthe sky could always use more than motes. more ambient life is on the table.

Devlog 20 - Three New Sets

this one adds the thing that’s been missing since the void set went in being actual custom armor that isn’t void. Luminous, Ash, and Null are new. 12 armor pieces, 15 tools, all with their own textures, their own powers, and their own place in the creative tab.

##the new sets

the void set already had its silhouette and its trim workflow. the new sets follow the same P-mode two-layer model so trims work the same, but they’re recolored palettes with their own read:- Luminous gold-cyan, glassy, the grove’s light. high enchantability, diamond-tier defense with a light-mobility kit.- Ash gray-ember with ember-orange highlights, volcano-forged. slightly lighter than void but built for fire.- Null violet-void, deleted matter. void-tier defense, hungrier enchant cost, gravity-deletion kit.

you can read them now

none of the three sets had working names before — datagen was fine, but the generated lang had Luminous_helmet with a capital L, so the game looked for item.endesium.luminous_helmet and found nothing. 26 keys showed item.endesium.___ in inventories.src/main/generated/assets/endesium/lang/en_us.json is now lowercase for all luminous/ash/null armor and tools, the phantom Ash_boots is gone. hover a luminous sword and it actually says luminous sword.

armor is per-piece now

new rule: every single piece does something on its own. the full set is an upgrade, not the only reason to wear it.Luminous — the grove’s light — GearAbilities per piece + tooltips on GearArmorItem:

• helmet Gleamsight – night vision while worn- chest Radiant Aegis — attackers get glowing 8s; full set Prism Ward adds darkness 3s + 3 magic burn-back- leggings Lightspeed — speed I (II with full set)- boots Lumen Leap — jump IIAsh — the volcano — volcano kit, fire is fuel:- helmet Ember Crown — fire resistance while worn- chest Searing Plate — attackers ignite 4s; burning/in lava gives strength I- leggings Magma Blood — regen I while burning/in lava- boots Ashwalker — stand on lava, and now hold shift in lava to sink at -0.18/tick to -0.55 like scaffolding, release and you stay down with LAVA puffs every 4tfull ash Volcanic Heart is permanent strength I. the set wants you in lava.Null — deleted matter — Erased Mind helmet purges levitation/darkness/nausea every tick, Erased Wound chest gives absorption 5s on hit, Weightless leggings no fall damage, Null Step boots +1 step height via attribute, full set Void Body deletes 25% of incoming projectiles with smoke + shield block feedback.

what’s next

more landmark variants, ambient life for the empty skies, and the deep end is still sitting there being ominous.

Sorry, no sneak peak yet - still not finished yet

Devlog #19

Sorry for the random numbering before - i guess i suck at counting. Anyways, devlog.

Updated structures

Before, structures were just decoration and weren’t technically vanilla structures (kaupenjoe misled me). However, I managed to turn them into structure with help from OpenCode. I also added some new structues, like a castle (needs some works still), some new cathedrals, willowtree trunk, etc. I am adding some more custom armor, and I will make it cool with cool abilities.

Custom armor

My plan for custom armor includes:

• Luminous armor (infinite night vision, light level control panel, more ideas to come),
• Ash armor (fire-related abilities, ashwalker boots already exist)
• Null armor (i have no idea on what to do for this, but it sounds cool)

Dragon fight

• Improve the dragon fight so it has several new stages
• Made it so that it spawns void wisps here and there
• Added a bunch of new drops it gives, working on adding Dragon Wings (slow but give iron-chestplate armor level)

What’s next?

• Add the custom armor and tools and weapons
• Finish the dragon fight so it is exceptional

Devlog 16

Short one, but it is the one I have been putting off. The dragon fight was authored with attacks, phases, telegraphs but the pillars were just decor and the last third of the health bar played identically to the first.

The pillars fight back

Crystal aegis is the simplest thing in the world: while crystals stand, thedragon heals. One line of health regen did more for the fight’s shape than anyattack I have authored, because it moves the win condition backwards. You donot win by out-DPS-ing a health bar; you win by climbing pillars under fire.Five crystals means two HP a second, which is enough to visibly undo a sloppyopening. That is the fight telling you it is watching.

Enrage is a ladder, not a switch

Three thresholds - 60, 35, 15 percent - each announced, each spawning void wisps into the arena. The wisps are not the threat; they are the clock. Every second you spend swatting a wisp off your back is a second the dragon is cycling toward catastrophe. From enrage two the dragon leaves breath pools under its flight path, which turns its flyovers into map control.

killing the dragon

The kill now drops a real pile: heart, fangs, bone, resonant scales. First kill still fires the transformation.
/dragonfight gives a live readout for tuning the curve without guessing.

Notes for next time

• The wisps cap at eight; if playtests say the floor is unreadable, that cap  is the first knob.
• Aegis regen is intentionally visible (1-2 HP/s). If it feels unfair, the  fix is fewer crystals required, not slower regen.
• Docs live in docs/DRAGON_FIGHT.md now, including the testing checklist.