Fantasma Cero
Protection against AI-LLM scrapers bots like the one of Cloudflare, But You Own the Data.
Description
Fantasma Cero is a minimalist and fast anti-AI/LLM WAF designed for those who need privacy and flexibility. Self-hosted, not a black-box.
Features
- Blocks known AI / LLM crawler behavior
- Optional JavaScript requirement
- Optional Proof-of-Work challenges
- Basic rate limiting
- Trap endpoints for bot detection
- Works behind Nginx or Caddy
- Runs as a single binary
- Stateless by default
- Designed for self-hosting
Installation
The following is a placeholder example. Replace with real instructions.
1. Download and install a release# set the proper release version
VERSION="v{latest_release}"
curl -L -O https://github.com/cocky-punch/fantasma_cero/releases/download/$VERSION/fantasma_cero-$VERSION-x86_64-unknown-linux-gnu.tar.gz
tar -xzf fantasma_cero-$VERSION-x86_64-unknown-linux-gnu.tar.gz
cd fantasma_cero-$VERSION-x86_64-unknown-linux-gnu
cp config.example.toml config.toml
nano config.toml
3. Check the examples of the config of webservers in ./examples for Nginx, Caddy.
Edit the one you need accordingly: port, host, etc; copy it into or merge with your Nginx or Caddy
config. Include it into your config.
bash
./fantasma_cero
Configuration
Configuration is file-based and intentionally small. Sensible defaults are provided.
# config.toml, essential settings
[server]
port = 8080
js_check_enabled = true
js_token_secret = "{{your secret}}"
#
# validation | validation_with_proxy
# "validation_with_proxy" = validate and proxy; used for debugging, testing
# validation - for production
#
operation_mode = "validation"
[pow_challenge]
# argon2 | sha256
algorithm = "sha256"
base_difficulty = 2
max_difficulty = 6
cookie_duration_days = 1
Project status
This is an early-stage open source project. Interfaces may change.
Contributions, issues, and design discussions are welcomed.