Start here

Gormes is a Go-native AI agent runtime: one static binary, no Python, no Docker, no Hermes process. Run it in your terminal as a TUI or as a persistent Telegram, Discord, or Slack gateway. Configuration and state live under ~/.gormes/. Secrets stay local.

60-second install

Pick one. Full details for each path live in Install.

Linux, macOS, WSL2:

curl -fsSLO https://github.com/TrebuchetDynamics/gormes-agent/releases/latest/download/install.sh
less install.sh
sh install.sh

Native Windows (PowerShell):

irm https://gormes.ai/install.ps1 -OutFile install.ps1
Get-Content .\install.ps1
powershell -ExecutionPolicy Bypass -File .\install.ps1

From source (requires Go 1.26+):

git clone https://github.com/TrebuchetDynamics/gormes-agent.git
cd gormes-agent
mkdir -p bin
CGO_ENABLED=0 go build -trimpath -o bin/gormes ./cmd/gormes
./bin/gormes --version
./bin/gormes doctor --offline

Verify the installer-published binary is on PATH:

gormes version
gormes doctor --offline

For source builds, keep using ./bin/gormes from the checkout, or add export PATH="$PWD/bin:$PATH" temporarily while testing that build.

Your first chat

Add a provider credential, then open chat.

gormes auth add openai --api-key sk-...
gormes chat

Or open the interactive TUI:

gormes

gormes auth add also accepts --type oauth for providers that support OAuth (Codex, Anthropic), plus --label, --inference-url, and credential type overrides. See gormes auth add --help for the full flag set.

For scripts, gormes chat -q "hello from Gormes" sends one query and exits without starting the TUI.

Now what?


Install	Linux/macOS, Windows, and source-build details
Configure	`config.toml`, environment variables, providers, Telegram, paths and logs
CLI reference	Every top-level command and subcommand
Recipes	First turn, Telegram bot, profiles, fallback chains, local Ollama
Troubleshooting	Doctor, common errors, log locations