MotelyJAML

The fastest Balatro seed searcher with JAML (or JSON) support and an interactive TUI.

Based on @tacodiva's incredible Motely - a blazing-fast SIMD-powered seed searcher. MotelyJAML extends it with multiple configuration formats (JAML, JSON) and a Terminal User Interface for easy filter creation.

Installation

📥 Download Latest Release

Available for Windows (x64), Linux (x64), macOS (Intel & Apple Silicon)

Build from Source

git clone https://github.com/OptimusPi/MotelyJAML.git
cd MotelyJAML
dotnet run --project Motely

Quick Start

Launch the TUI (Terminal User Interface)

# Launch interactive TUI (default when no arguments provided)
dotnet run

The TUI provides an interactive menu for:

• Building custom filters
  • It's really simple, supports all item types
  • TODO: add support for edition, stickers/sealts, etc.
• Loading config files and searching
  • P.S. Sorry about the BalatroShaderBackground, I just had to
  • shader background PAUSES during API Host or Search
• Starting the API server
  • Optional:Install cloudflared
  • Navigate to SETTINGS and change search/api as needed.
  • Host
    • set the host. usually 127.0.0.1 or localhost
    • host is sometimes + on some linux distros/VMs.
    • TODO: I'm new to Apple so still learning pitfalls of local networking?
    • set the port to whatever you want.
  • Search
    • lower the search to avoid 100% CPU load
    • Example: if you have a 16-core CPU but you are playing a game, doing homework, browsing the web, you could set --threads 8 to save a lot of system resources for yourself.
    • Example: if you have a really tiny box to put this on, maybe it needs less threads to save a little wiggle room for SSH/VNC/Parsec/Steamlink/ to still function with good performance
    • Example: if you have an 8 core CPU but want 2 different search instances running you can run 2 different terminals each with -- threads 4
    • NOTE: Never set the threads past what your CPU has. You'll lower performance.
    • NOTE: if your filter is very permissive, you'll bottleneck on console.writeline()
    • NOTE: see below for more options;

Command Line Usage

Search with JSON/JAML filters

dotnet run -- --json PerkeoObservatory --threads 16 --cutoff 2
dotnet run -- --jaml MyFilter --threads 16 --cutoff 2

Use Native C# Filters

Built-in filters from tacodiva's original Motely:

dotnet run -- --native PerkeoObservatory --threads 16

Analyze Seeds

Following the analyzer.cl convention from Immolate, TheSoul, and Blueprint:

dotnet run -- --analyze 5SC1HR14

Available Options

Core:

• --json <name> - Load filter from JsonFilters/
• --jaml <name> - Load filter from JamlFilters/
• --native <name> - Use built-in C# filter
• --analyze <SEED> - Analyze specific seed
• (no args) - Launch TUI

Performance:

• --threads <N> - Thread count (default: CPU cores)
• --batchSize <1-8> - Batch size for vectorization
• --startBatch <N> / --endBatch <N> - Search range

Filtering:

• --cutoff <N|auto> - Minimum score (JSON/JAML only)
• --deck <DECK> - Override filter's deck
• --stake <STAKE> - Override filter's stake

Filter Formats

JSON Filter Format

Create in JsonFilters/:

{
  "name": "Example",
  "must": [{
    "type": "Voucher",
    "value": "Telescope",
    "antes": [1, 2, 3]
  }],
  "should": [{
    "type": "Joker",
    "value": "Blueprint",
    "antes": [1, 2, 3, 4],
    "score": 100
  }]
}

JAML Filter Format

Create in JamlFilters/:

name: Example
description: Example filter using JAML
author: YourName
dateCreated: 2025-01-01T00:00:00Z

must:
  - type: Voucher
    value: Telescope
    antes: [1, 2, 3]

should:
  - type: Joker
    value: Blueprint
    antes: [1, 2, 3, 4]
    score: 100

Both formats support the same filter logic - choose whichever you prefer!

Using Defaults in JAML

Configure defaults for your filter to avoid repetition:

name: My Filter
defaults:
  # Default antes (1-8)
  antes: [1, 2, 3, 4, 5, 6, 7, 8]

  # Default pack/shop slots (auto-adjusted for ante 1)
  packSlots: [0, 1, 2, 3, 4, 5]
  shopSlots: [0, 1, 2, 3, 4, 5]

  # Default score
  score: 1

must:
  # Uses defaults.antes automatically
  - voucher: Telescope

  # Override with specific antes
  - voucher: Observatory
    antes: [2, 3]

should:
  # Uses defaults.score = 1
  - joker: Blueprint

  # Override score
  - joker: Brainstorm
    score: 50

Note: Ante 1 automatically limits pack slots to [0-3] and shop slots to [0-3] (4 slots max in ante 1).

Native Filters

• ErraticFinder: Erratic Deck with 32+ cards of a specific suit
• negativecopy: Showman + copy jokers with negatives
• PerkeoObservatory: Telescope/Observatory + soul jokers
• trickeoglyph: Cartomancer + Hieroglyph

Advanced Filter Features

Alternative Joker Sources

Beyond standard shop and pack slots, jokers can also come from special sources:

must:
  - type: Joker
    value: Blueprint
    antes: [1]
    sources:
      judgement: [0]      # First Judgement tarot usage
      rareTag: [0]        # First Rare Tag
      uncommonTag: [0, 1] # First two Uncommon Tags

Available Sources:

• shopSlots: Shop item slots (default: auto-detected per ante)
• packSlots: Booster pack slots (default: 0-5)
• judgement: Judgement tarot card joker rolls (gives random joker)
• rareTag: Rare Tag joker rolls (guarantees rare joker)
• uncommonTag: Uncommon Tag joker rolls (guarantees uncommon joker)

Event Filters

Filter seeds by specific random event outcomes:

must:
  - type: Event
    eventType: LuckyMoney
    antes: [1, 2]
    rolls: [0, 1]  # Check first two Lucky money triggers
    score: 10

Available Event Types:

• LuckyMoney - Lucky enhancement triggers $1 money drop
• LuckyMult - Lucky enhancement triggers +mult bonus
• MisprintMult - Misprint joker mult value roll
• WheelOfFortune - Wheel of Fortune tarot edition roll
• CavendishExtinct - Cavendish banana extinction check
• GrosMichelExtinct - Gros Michel banana extinction check

The rolls array specifies which roll indices to check (0-indexed). For example, rolls: [0, 1] checks the first and second occurrences of that event.

Tweak the Batch Size

1. --batchSize 1: 35^1 = 35 seeds per batch
2. --batchSize 2: 35^2 = 1,225 seeds per batch
3. --batchSize 3: 35^3 = 42,875 seeds per batch
4. --batchSize 4: 35^4 = 1,500,625 seeds per batch
5. --batchSize 5: 35^5 = 52,521,875 seeds per batch
6. --batchSize 6: 35^6 = 1,838,265,625 seeds per batch
7. --batchSize 7: 35^7 = 64,339,296,875 seeds per batch

Higher batch sizes = less responsive updates but faster overall search. Recommended: 2-4 for interactive use, 4-5 for overnight searches.