Skip to main content

Erlang on Miren

Erlang isn't auto-detected, so you deploy it with a Dockerfile.miren that builds a rebar3 release and runs it on the BEAM. This guide uses Cowboy as the HTTP server. (For Elixir and Gleam, which also run on the BEAM, see their own guides: Elixir, Gleam.)

Let your agent do this

Ask your AI coding agent to "set up this Erlang app on Miren" after installing the Miren agent skills. It adds the Dockerfile.miren, binds Cowboy to 0.0.0.0:$PORT, and deploys — using this page as its reference.

Does this source build need a Dockerfile?​

Yes. Miren doesn't auto-detect Erlang, so add a Dockerfile.miren to your project root. Miren builds from it instead of guessing the stack — see Using Dockerfile.miren.

Want native support?

Miren auto-detects and builds common stacks (Python, Node, Bun, Go, Ruby, Rust) without a Dockerfile. This language isn't one of them yet — if you'd like first-class support, request it.

Read the injected port​

Miren injects PORT and routes traffic to it. In the application's start/2, read PORT and start Cowboy's listener on it (Cowboy binds all interfaces by default):

-module(erlang_bench_app).
-behaviour(application).
-export([start/2, stop/1]).

start(_Type, _Args) ->
    Dispatch = cowboy_router:compile([
        {'_', [{"/", erlang_bench_handler, []}]}
    ]),
    Port = list_to_integer(os:getenv("PORT", "8080")),
    {ok, _} = cowboy:start_clear(http_listener,
        [{port, Port}],
        #{env => #{dispatch => Dispatch}}),
    erlang_bench_sup:start_link().

stop(_State) -> ok.

A minimal request handler:

-module(erlang_bench_handler).
-export([init/2]).

init(Req0, State) ->
    Req = cowboy_req:reply(200,
        #{<<"content-type">> => <<"text/plain">>},
        <<"Hello from Erlang on Miren!\n">>,
        Req0),
    {ok, Req, State}.

Declare Cowboy in rebar.config and the release:

{deps, [{cowboy, "2.12.0"}]}.
{relx, [{release, {erlang_bench, "0.1.0"}, [erlang_bench, cowboy]},
        {dev_mode, false},
        {include_erts, true}]}.

include_erts, true bundles the runtime, so the release is self-contained.

The Dockerfile​

Create Dockerfile.miren in your project root. The build produces a release; the runtime image just needs a couple of shared libraries the bundled ERTS links against:

# ----- Build stage -----
FROM erlang:27 AS builder
WORKDIR /app
COPY . .
RUN rebar3 release

# ----- Runtime stage -----
FROM debian:12-slim
RUN apt-get update -y && apt-get install -y libncurses6 libssl3 && rm -rf /var/lib/apt/lists/*
COPY --from=builder /app/_build/default/rel/erlang_bench /app
EXPOSE 8080
CMD ["/app/bin/erlang_bench", "foreground"]

.dockerignore​

.git
_build

Deploy​

The Dockerfile's CMD starts the app. Miren uses it as the web service's startup default, so you don't need a Procfile or service command.

Create .miren/app.toml naming your app and deploy from your project root:

name = "erlang-bench"
miren deploy

Environment variables​

Set variables with miren env set — -e for plain values, -s for secrets (masked in output and logs). Read them with os:getenv("KEY"):

miren env set -e LOG_LEVEL=info
miren env set -s DATABASE_URL

You can also declare variables in .miren/app.toml:

[[env]]
key = "DATABASE_URL"
value = ""
required = true
sensitive = true

See App Configuration — Environment Variables.

Agent quick reference​

  • Detection: none — requires Dockerfile.miren (rebar3 release)
  • Build: rebar3 release on erlang:27; include_erts, true makes it self-contained
  • Runtime libs: libncurses6 libssl3 on debian-slim
  • Startup: inherited from the Dockerfile CMD; no Procfile or service command needed
  • Port: os:getenv("PORT", "8080"); cowboy:start_clear(_, [{port, Port}], _)
  • Env vars: miren env set -e/-s; read with os:getenv

Next steps​