Differences

This shows you the differences between two versions of the page.

--- ai:atlas-prompts [2026/04/15 23:23] – [🎨 Abstract Emotion] mh
+++ ai:atlas-prompts [2026/07/15 22:43] (current) – [Workflow] mh
@@ Line 322: / Line 322: @@
 ----
-===== 🧬 Non-Human & Creature Logic =====
+===== 🧬 Mythic =====
-==== 🛡️ Fantasy Humanoid Character ====
+==== 🛡️ Fantasy Warrior ====
 <code txt>
@@ Line 331: / Line 331: @@
 👉 Tests: layering logic, costume complexity, material differentiation, humanoid design coherence
+----
+==== 🧙 Mythic Caster / Mage ====
+<code txt>
+a powerful female mage standing in a mystical environment, flowing garments, magical energy emanating from hands, subtle arcane effects, grounded realism
+</code>
+👉 Tests: energy effects, pose + gesture coherence, integration of magic with realism
 ----
@@ Line 353: / Line 363: @@
 ----
 ===== 🏙️ Environment & Scale =====
@@ Line 512: / Line 523: @@
 ----
-===== 🧩 Precision & Constraint =====
+==== 💥 Material Impact ====
-==== 🏛️ Perfect Symmetry ====
+<code txt>
+a solid object breaking apart on impact, fragments and debris flying, realistic material fracture, frozen motion, high detail
+</code>
+👉 Tests: rigid body physics, fragmentation logic, motion coherence, debris distribution
+===== 🧩 Structure & Symbols =====
+==== 🏛️ Architectural Symmetry ====
 <code txt>
@@ Line 531: / Line 550: @@
 👉 Tests: pattern consistency, detail repetition, drift resistance
+----
+==== 🔤 Embedded Readable Text ====
+<code txt>
+a sign or surface with clearly visible English text reading "The Old Library", integrated naturally into a scene, realistic materials, readable typography, natural lighting
+</code>
+👉 Tests: spelling accuracy, text legibility, typography coherence, integration of text into environment
+----
+==== 🔮 Fantasy Runes & Glyphs ====
+<code txt>
+an ancient surface covered in glowing magical runes and glyphs, intricate symbols, non-human writing system, carved or engraved, mystical atmosphere, soft ambient light
+</code>
+👉 Tests: symbolic consistency, pseudo-language generation, pattern logic, stylization of non-human writing systems
 ----
@@ Line 576: / Line 615: @@
 ----
-===== ✍️ Writing Systems =====
-==== 🔤 Embedded Readable Text ====
+===== Local AI add-ons =====
+==== Positive Baseline (SDXL) ====
 <code txt>
-a sign or surface with clearly visible English text reading "The Old Library", integrated naturally into a scene, realistic materials, readable typography, natural lighting
+high quality image, detailed, coherent composition, <REST OF PROMPT>
 </code>
-👉 Tests: spelling accuracy, text legibility, typography coherence, integration of text into environment
+==== Negative Baseline ====
-----
+<code txt>
+low quality, worst quality, low resolution, blurry, jpeg artifacts, watermark, signature, logo, unwanted text, text artifacts, bad anatomy, distorted anatomy, deformed body, extra limbs, missing limbs, extra fingers, missing fingers, fused fingers, malformed hands, malformed feet, duplicate subject, cropped subject, out of frame
+</code>
-==== 🔮 Fantasy Runes & Glyphs ====
+==== Workflow ====
+Lock the whole ritual circle:
 <code txt>
-an ancient surface covered in glowing magical runes and glyphs, intricate symbols, non-human writing system, carved or engraved, mystical atmosphere, soft ambient light
+same prompt
+same negative prompt
+same checkpoint type family
+same resolution
+same seed policy
+same sampler
+same scheduler
+same steps
+same CFG
+same VAE behavior
+no LoRA
+no ControlNet
+no IPAdapter
+no ref image
+no upscale
+no inpainting
 </code>
-👉 Tests: symbolic consistency, pseudo-language generation, pattern logic, stylization of non-human writing systems
+At first: checkpoints only.
-----
+Then later:
+<code txt>
+checkpoint + LoRA
+checkpoint + sampler variation
+checkpoint + CFG variation
+checkpoint + style LoRA
+checkpoint + ControlNet
+</code>
+Don’t mix those into the baseline.
+Baseline settings suggestion for SDXL checkpoints:
+<code txt>
+Resolution: 1024 x 1024 for square prompts
+Steps: 30
+CFG: 6
+Sampler: dpmpp_2m
+Scheduler: karras
+Batch size: 1
+Seed: fixed per prompt across all checkpoints
+LoRA: none
+Refiner: none, unless every checkpoint test uses the same refiner strategy
+</code>
+For non-square prompts, use SDXL-friendly sizes close to 1 megapixel:
+<code txt>
+Portrait: 832 x 1216
+Landscape: 1216 x 832
+Wide: 1344 x 768
+Square: 1024 x 1024
+</code>