Skip to content

Technical guide

FlameCord vs VeloFlame: pick by proxy ecosystem, not speed

FlameCord is a BungeeCord/Waterfall fork and VeloFlame a Velocity fork, so compare protocol tables, build targets, firewall defaults and migration limits.

Products and contents

FlameCord and VeloFlame are different proxies, not two builds of the same one. FlameCord is a BungeeCord/Waterfall fork. VeloFlame is a Velocity fork built from the Velocity-CTD tree. Your plugins, forwarding setup and staff commands should decide the move. Neither repository has a benchmark, so “faster” is not a reason to switch.

They fork different proxies

FlameCord’s README calls it a fork of BungeeCord and Waterfall. Its proxy module inherits the Waterfall 26.1-R0.1-SNAPSHOT line, and the bootstrap main class is still net.md_5.bungee.Bootstrap. Bungee plugins, Bungee-style commands and Bungee forwarding are native to it.

VeloFlame’s root project is named velocity-ctd. Its product code adds /veloflame commands and its own firewall. Velocity plugins and Velocity forwarding are native. A Bungee plugin jar won’t load on VeloFlame, and a Velocity plugin won’t load on FlameCord.

Protocol and build facts

Fact FlameCord VeloFlame
Upstream BungeeCord / Waterfall Velocity-CTD
Protocol table 1.8.x through 1.21.x, plus a 26.x line ending at MINECRAFT_26_3 minimum 1.7.2, maximum constant MINECRAFT_26_3
Compile target proxy release 17, bootstrap release 11 Java 25 toolchain
Product version string Waterfall parent 26.1-R0.1-SNAPSHOT Gradle 4.2.1-SNAPSHOT, internal string 1.4.9
Shipped artifact FlameCord.jar from the bootstrap module VeloFlame-Lite.jar from the shaded proxy module
Product commands none registered by the FlameCord module; upstream module toggles remain /veloflame reload, whitelist, firewall, maintenance
Config files flamecord.yml, modules.yml veloflame.toml, default-velocity.toml

These are manifest and protocol-table facts, not runtime checks. The compile targets show how the jars are built. Neither repository states a runtime Java requirement in its product docs. Don’t treat release 17 or Java 25 as a hosting requirement until you test your own startup.

Both protocol tables are broad, but a client version in a protocol enum only proves the proxy can parse that handshake. It doesn’t prove gameplay works.

Firewall defaults and root requirements

Both proxies ship with an IP firewall on by default. Both expect Linux tools that ordinary panel hosting may not provide.

FlameCord’s antibot.firewall.ipset defaults to true. At startup, the code runs apt install iptables, apt install ipset, ipset create flamecord hash:ip timeout <expiry>, and adds an iptables -t raw -I PREROUTING drop rule. It removes the rule on shutdown. This needs a Debian-style host and root. The repository doesn’t say what happens when these commands fail.

VeloFlame’s firewall.enabled defaults to true, with firewall.ipset = true and a 60-second duration. The shipped config comment names the exact iptables -w 5 -I INPUT rule and the requirement: Linux ipset and iptables, plus root or CAP_NET_ADMIN. It also says the Java-side firewall cache still works if OS setup fails.

Decide before you migrate whether you can run that tooling. If you can’t, turn the feature off on purpose and accept the weaker protection. Don’t leave a default that fails silently.

Artifacts and version drift

Don’t copy a jar name from a README.

  • FlameCord builds FlameCord.jar (bootstrap finalName), not a Waterfall-named file.
  • VeloFlame’s README is upstream Velocity-CTD text and mentions an “-all JAR”. The real shadow jar is hard-coded as VeloFlame-Lite.jar.
  • VeloFlame reports two version strings: the Gradle version and internal 1.4.9. Include /veloflame output when you ask for support.

After a build, list the output directory and run the jar you actually built. For deployment, see the FlameCord product page and docs, or VeloFlame and its docs.

Migration constraints

  1. Plugins. List every proxy plugin. Replace Bungee plugins with Velocity equivalents. Don’t port them by renaming the jar.
  2. Forwarding. Bungee-style ip_forward and the backend’s Bungee mode don’t carry over to Velocity’s player-info-forwarding-mode and forwarding-secret-file. VeloFlame ships with player-info-forwarding-mode = "none", so the default config is not a finished production setup.
  3. Commands and permissions. FlameCord has no product command tree. VeloFlame adds /veloflame with veloflame.reload, veloflame.whitelist, veloflame.firewall and veloflame.maintenance. Rewrite staff docs and permission grants.
  4. Configs. There is no converter. Rebuild veloflame.toml and default-velocity.toml from the new defaults, then reapply only the settings you can explain.
  5. Cutover. Run both proxies against a test backend. Compare login, forwarding, whitelist and firewall behavior. Then change DNS. A proxy swap is not a rolling restart.

The secure proxy and backend guide covers the forwarding and trust layer that has to survive the move.

Frequently asked questions

Can I run FlameCord on Velocity?

No. FlameCord is a BungeeCord and Waterfall fork. VeloFlame is the Velocity fork.

How should I choose between them?

Choose by proxy ecosystem. Your plugins, forwarding setup and staff commands should decide the move.

Does FlameCord need root?

Its default firewall module runs ipset and iptables, so it needs a Linux host with root. The firewall is enabled by default.