Tower of Hell section filtering is reported by community guides as a way to narrow private-server section selection by properties. Current references mention tags such as conveyor, kills, locals, and phantom, plus ! and & operators. Other player reports say the feature stopped working or changed. Treat every expression as experimental until the live owner interface confirms it.
Understand the evidence limit
The official experience description confirms VIP settings for skipping, tower size, and shop locking. It does not publish section-filter syntax. A community wiki can preserve a real older feature and still be wrong for the current client. That disagreement is why this page does not promise a working command or a complete tag set.
Do not buy a private server solely for filtering unless you can see or reproduce the current control.
Interpret reported tags cautiously
Community documentation uses conveyor for sections with conveyor behavior, kills for killpart-related sections, locals for locally acting content, and phantom for a particular property or category. The exact definitions and pool coverage are not official. A section can combine multiple properties, and an updated tag may classify it differently.
Use a tag as a hypothesis: “Does the resulting tower contain the reported property?” Record results rather than assuming the label is correct.
Interpret reported operators
The reported ! operator is commonly presented as exclusion, while & is presented as combining requirements. The position, spacing, and complete expression format are not safely established by first-party documentation. Do not invent a command wrapper around these symbols.
If the live settings field provides examples or validation, copy that format exactly. If it does not, test only a simple non-destructive expression and observe the response.
Run a reproducible test
- Use an official private server you are authorized to control.
- Record Place ID, date, tower size, and all starting settings.
- Look for a visible filter field or current help response.
- Enter one simple supported-looking tag exactly once.
- Start a fresh tower using the normal control.
- Record every resulting section and whether the target property appears.
- Repeat across several towers before drawing a conclusion.
- Clear the filter and verify ordinary selection returns.
One matching section can occur randomly and does not prove the filter worked. A visible success response plus repeatable pool change is stronger evidence.
Test exclusion and combinations later
Only after a single tag works should you test exclusion or intersection. Hold tower size and other settings constant. Compare the result with a baseline set collected without filtering. If syntax is rejected or results remain indistinguishable, label the feature unresolved instead of cycling commands indefinitely.
Do not combine filters with mass skipping in a shared session without warning players.
Use filtering for focused practice
If verified, a conveyor filter could provide more repetitions on forced movement, while a killpart filter could support clearance and timing practice. The goal is skill transfer. After focused work, return to unfiltered random towers and confirm that the technique survives ordinary combinations.
A filter should not become a claim that one listed section always belongs to one tag. Maintain the observation date and current place.
Troubleshoot no effect
Confirm owner permission, current private server, correct field or command context, and whether the setting applies next round. Remove extra spaces or arguments only if the live help indicates their format. Rejoin once if the UI appears stale.
If current reports conflict and no official control appears, stop. Treat filtering as unavailable or unverified for the session. Do not use an executor, altered client, or browser script to force it.
Evidence checklist
- Official universe and authorized private server confirmed.
- Visible current control or response captured.
- Baseline tower set recorded.
- One tag tested before operators.
- Other settings held constant.
- Several results observed.
- Random matches not treated as proof.
- Filter cleared afterward.
- Version date attached to every conclusion.
Sources and confidence
Tag names and operators are community-reported details reviewed August 23, 2026. Community reports conflict about whether filtering remains supported. The official description does not document the syntax. Current live owner controls, validation messages, and reproducible section results are the only basis for a present-tense claim.