Skip to main content

cc-liquid - open source reference implementation for a simple, metamodel-based Hyperliquid portfolio rebalancer

Project description

cc-liquid

⚠️ PRE-ALPHA SOFTWARE - USE AT YOUR OWN RISK ⚠️

This is PRE-ALPHA software provided as a reference implementation only. By using this software, you acknowledge:

  • Risk of COMPLETE LOSS of funds - CrowdCent makes NO WARRANTIES and assumes NO LIABILITY
  • Software is provided "AS IS" without any guarantees
  • You must comply with all Hyperliquid terms of service
  • We do NOT endorse any vaults, strategies, or implementations using this tool

Overview

cc-liquid is a simple metamodel-based rebalancer for Hyperliquid that:

  • Downloads metamodel predictions from CrowdCent or Numerai
  • Identifies top long and bottom short opportunities
  • Rebalances your Hyperliquid portfolio based on predictions
  • Manages position sizing and risk automatically based on portfolio configurations

Read the docs: documentation

dashboard

Installation

# Install globally
uv tool install cc-liquid

# With Numerai support
uv tool install cc-liquid[numerai]

# Or run without installing
uvx cc-liquid --help

# Optional: Enable shell completion
cc-liquid completion install

Quick Start

cc-liquid init        # Interactive setup wizard (recommended for new users)

The wizard guides you through:

  • Choosing testnet (recommended) or mainnet
  • Selecting data source (CrowdCent/Numerai/local)
  • Entering API keys securely
  • Setting portfolio parameters
  • Auto-configuring for safety

Configuration

Secrets (.env only - never commit)

# API Keys
CROWDCENT_API_KEY=...

# Signer private keys (owner key or approved agent wallet keys)
HYPERLIQUID_PRIVATE_KEY=0x...
HYPER_AGENT_KEY_PERSONAL=0x...
HYPER_AGENT_KEY_VAULT=0x...

Get your keys from:

  • CrowdCent: https://crowdcent.com/profile
  • Hyperliquid: https://app.hyperliquid.xyz/API

Settings (cc-liquid-config.yaml)

profiles:
  default:
    owner: 0xYourMain           # Main owner address (never the API/Agent wallet)
    vault: null                 # null for personal, or vault address
    signer_env: HYPERLIQUID_PRIVATE_KEY  # ENV var name (not the key itself)

  my-vault:
    owner: 0xYourMain           # optional, informational
    vault: 0xVaultAddress
    signer_env: HYPER_AGENT_KEY_VAULT

active_profile: default

data:
  source: crowdcent             # crowdcent, numerai, or local
  path: predictions.parquet

portfolio:
  num_long: 10
  num_short: 10
  target_leverage: 1.0          # 1.0 = no leverage, 2.0 = 2x, etc.

execution:
  slippage_tolerance: 0.005
  min_trade_value: 10.0

Profile Management:

cc-liquid profile list
cc-liquid profile show
cc-liquid profile use <name>

Agent Wallets (Recommended for automation):

  • Use separate agent keys per bot for nonce isolation
  • Approve agent in Hyperliquid UI, then add private key to .env
  • Never use agent address for Info queries (use owner/vault)

Usage

Basic Commands

# Account info
cc-liquid account

# Download predictions
cc-liquid download-crowdcent -o predictions.parquet
cc-liquid download-numerai -o predictions.parquet

# Rebalance portfolio
cc-liquid rebalance

# Continuous rebalancing (WARNING: auto-executes)
cc-liquid run --skip-confirm

Configuration Overrides

Use --set to override any config value at runtime:

# Data source (auto-applies column mappings)
--set data.source=numerai      # → date, symbol, meta_model columns
--set data.source=crowdcent    # → release_date, id columns
--set data.path=custom.parquet

# Portfolio
--set portfolio.num_long=15
--set portfolio.num_short=8
--set portfolio.target_leverage=2.0

# Execution
--set execution.slippage_tolerance=0.01

# Environment
--set is_testnet=true

Example combinations:

cc-liquid rebalance --set data.source=numerai --set portfolio.num_long=20 --set portfolio.target_leverage=2.0

How It Works

  1. Download Metamodel: Fetches consolidated predictions from CrowdCent or Numerai
  2. Select Assets: Identifies top N longs and bottom N shorts based on 10-day predictions
  3. Calculate Positions: Determines target sizes based on account value and leverage
  4. Generate Trades: Calculates required trades from current to target positions
  5. Execute Orders: Places market orders with configured slippage tolerance

Backtesting & Optimization

cc-liquid includes backtesting and optimization tools to test strategies on historical data:

cc-liquid analyze                    # Run backtest with current config
cc-liquid optimize                   # Find optimal parameters via grid search

⚠️ BACKTESTING DISCLAIMER ⚠️

Past performance does not predict future results. Backtesting has inherent limitations:

  • Results are hypothetical and may not reflect real trading conditions
  • Optimized parameters are prone to overfitting
  • Market conditions, liquidity, and execution costs change over time
  • Always validate with out-of-sample data and paper trading before live deployment

See full documentation for details.

Future Enhancements

  • Neutralize to known risk factors
  • Unequal-weight position sizing
  • Advanced portfolio optimization

License

MIT License - See LICENSE file for details

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cc_liquid-0.1.4a4.tar.gz (589.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

cc_liquid-0.1.4a4-py3-none-any.whl (60.9 kB view details)

Uploaded Python 3

File details

Details for the file cc_liquid-0.1.4a4.tar.gz.

File metadata

  • Download URL: cc_liquid-0.1.4a4.tar.gz
  • Upload date:
  • Size: 589.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.7.13

File hashes

Hashes for cc_liquid-0.1.4a4.tar.gz
Algorithm Hash digest
SHA256 66570225c9b788dcd853f581310d25dafa07e2bc4beb66e1668d5c6350fa4aed
MD5 4ee3f70f10f77b1a5d09c2c3f9f9add9
BLAKE2b-256 8275cb029553a8815ec767c44bd80bbb2b6fd9ad635a7fbc0f1f338892bf610c

See more details on using hashes here.

File details

Details for the file cc_liquid-0.1.4a4-py3-none-any.whl.

File metadata

File hashes

Hashes for cc_liquid-0.1.4a4-py3-none-any.whl
Algorithm Hash digest
SHA256 ad1cead912b44059f799f4f339afaee44f73a8bb527783246a849eac7b56b411
MD5 ddc5b5426b945efb92a619e1007eddd7
BLAKE2b-256 8928acdabcc48aa7afbf10747747d949cbcbcc0676feed950551053169991f65

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page