gormes-agent
Health Warn
- No license — Repository has no license file
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Low visibility — Only 5 GitHub stars
Code Pass
- Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
No AI report is available for this listing yet.
The agent that GOes with you
Hermes-class agents from one Go binary.
Gormes is a Go-native runtime for Hermes-compatible agents: providers, tools, skills, local SQLite memory, sessions, and gateways in one static binary. It carries the 30 most-used Hermes skills to Termux, Windows-without-Python, and locked-down corp Linux — no pip, no venv, no Docker daemon.
Gormes is not a micro-agent. It keeps the broad Hermes agent architecture and makes it portable, inspectable, and cheap to operate from a normal terminal.
At A Glance
| Signal | Current evidence |
|---|---|
| Runtime shape | One Go binary for CLI, TUI, provider turns, tools, skills, memory, sessions, dashboard, and gateways |
| Install proof | gormes doctor --offline and gormes --offline run before any provider token is needed |
| Measured footprint | Linux build ~45.8 MB; offline doctor peak RSS ~25.6 MB (benchmarks.json, 2026-05-16) |
| Local state | SQLite under ~/.gormes; no Redis, vector DB, Python service, or Node service on the local path |
| Stable channels | Telegram, Discord, and Slack through one gateway process |
| Project posture | Scout release: useful today, with full Hermes parity, voice/TTS, signing, and package-manager lanes still hardening |
Quick Install
Linux, macOS, WSL2:
curl -fsSL https://github.com/TrebuchetDynamics/gormes-agent/releases/latest/download/install.sh | sh
Windows (native PowerShell):
irm https://raw.githubusercontent.com/TrebuchetDynamics/gormes-agent/main/scripts/install.ps1 | iex
Inspect first: download the script, read it, then run it. Both installers are user-scoped — no root or admin paths.
After installation:
gormes doctor --offline
gormes setup
gormes chat
That is the shortest path: prove the runtime locally, configure provider/model, then open a provider-backed terminal chat.
First Setup
gormes setup # guided setup for provider, model, terminal, gateway, tools
gormes setup provider # direct provider setup shortcut
gormes chat # provider-backed terminal chat
gormes doctor --offline # local runtime, TUI, gateway, memory — no credentials needed
gormes --offline # native TUI, no network
If gormes chat opens, the TUI and gateway have a model to use.
Built For
- Developers who want a real agent runtime without Python environment drift.
- Operators who need offline diagnostics before adding provider credentials.
- Small servers, Termux/Android, WSL2, and locked-down Linux hosts where Docker or venv repair is friction.
- Termux can be the controller while a remote SSH host handles Docker, browser automation, GPU/local models, and large builds.
- Long-running personal or team agents that need local sessions, memory, tools, and chat gateways.
Not yet for teams that require signed enterprise releases, full Hermes parity, voice/TTS parity, or broad channel parity on day one.
What Works Today
| Surface | Status |
|---|---|
| Install, offline smoke, doctor, dashboard | Supported |
| CLI-first setup/config, TUI, scripted chat, multi-agent routing | Supported |
| Providers: OpenAI, Anthropic, DeepSeek, Groq, Ollama, OpenAI Codex, OpenCode, custom endpoints | Supported |
| Local SQLite memory (Goncho), session state | Supported |
| Gateways: Telegram, Discord, Slack | Supported |
| Profiles, local Kanban board, skills/plugins inventory, security/secret audits | Supported |
| Hermes / OpenClaw migration with dry-run | Supported |
| Gateways: WhatsApp, Teams, Yuanbao | Experimental |
| Learning loop, MCP/plugin parity, voice/TTS | Roadmap |
| Release signing, package-manager lanes | Roadmap |
Full Hermes-parity status by phase lives in the roadmap.
Why People Switch
| Other agents | Gormes | |
|---|---|---|
| Install | pip, venvs, system packages | One Go command from install.sh |
| First setup | Find and edit config files | gormes setup |
| Smoke test | Needs a live model first | gormes doctor --offline and gormes --offline |
| State | Redis, vector DBs, sidecars | SQLite under ~/.gormes |
| Channels | Separate bot glue per platform | One gateway process with channel bindings |
| Footprint claims | Often anecdotal | Binary size and RSS recorded in benchmarks.json |
| Release trust | Ad-hoc local environments | Tagged release assets with SHA-256 + SBOMs |
How It Works
Gormes follows the Hermes agent shape, but moves the operational surface into Go:
cmd/gormesowns the CLI, setup, TUI entry point, dashboard, and gateway commands.internal/kernelruns the turn loop shared by the TUI, one-shot chat, and channel gateway.internal/providerandinternal/hermesadapt OpenAI-compatible, Anthropic, DeepSeek, Groq, Ollama, Codex, OpenCode, and custom endpoints.internal/toolsandinternal/skillsexpose the tool and skill registry without a Python sidecar.internal/goncho,internal/memory, andinternal/sessionkeep local memory and session state inspectable in SQLite.
Daily Use
gormes # open the native TUI
gormes setup # guided setup and reconfiguration
gormes setup --quick # fill missing setup items only
gormes dashboard # web UI at http://127.0.0.1:43827/dashboard
gormes config show # inspect config with secrets redacted
gormes profile use <name> # switch isolated profile homes
gormes gateway # run the configured messaging gateway
gormes gateway status --json # inspect gateway runtime state
gormes logs # read recent gateway logs
CLI vs Gateway
| Action | Local CLI | Messaging gateway |
|---|---|---|
| Start chatting | gormes |
Run gormes gateway, then message the configured channel bot. |
| Start provider-backed chat | gormes chat |
Send a normal message after provider setup. |
| Configure provider/model | gormes setup provider |
Run setup locally, then gormes gateway reload. |
| Diagnose runtime issues | gormes doctor --offline, gormes logs |
gormes gateway status, then fix config or credentials locally. |
Migrating From Hermes Or OpenClaw
Always dry-run first and review the redacted manifest before applying.
gormes migrate hermes --dry-run --source ~/.hermes
gormes migrate hermes --yes --source ~/.hermes
gormes migrate openclaw --dry-run --source ~/.openclaw
gormes migrate openclaw --yes --source ~/.openclaw
gormes migrate openclaw --yes --secrets --source ~/.openclaw
OpenClaw cleanup archives old directories rather than deleting them:
gormes migrate openclaw cleanup --dry-run
gormes migrate openclaw cleanup --yes
Build From Source
git clone https://github.com/TrebuchetDynamics/gormes-agent.git
cd gormes-agent
make build
export PATH="$PWD/bin:$PATH"
gormes doctor --offline
Docs
| Section | What it covers |
|---|---|
| Installation | Source build, installer path, PATH, and offline verification. |
| First run | doctor, setup, provider setup, and first provider-backed turn. |
| What you can do | Outcome-driven recipes for CLI, config, gateway, profiles, memory, and security. |
| CLI reference | Commands, flags, and operator workflows. |
| Providers | Supported provider config and credential paths. |
| Configuration | ~/.gormes/config.toml, .env, agents, workspaces, and bindings. |
| Gateway | Channel runtime, status checks, reloads, and troubleshooting evidence. |
| Troubleshooting | Common install, provider, gateway, and browser-tool failures. |
| Roadmap | Hermes parity phases, status labels, and progress evidence. |
Security & Trust
gormes doctor --offlineandgormes --offlineprove local readiness before any token spend.- Provider secrets stay local under
~/.gormes/.env; config under~/.gormes/config.toml. - Tagged releases publish a single static binary per target plus SHA-256 checksums and SBOMs (release signing and package-manager lanes are still hardening).
- The
curl … | shinstall path is validated by an end-to-end suite (tests/install/e2e.sh) covering API outage with redirect fallback, SHA-256 mismatch abort, SSH-origin update fallback to public HTTPS, hosts without Go/curl/wget/systemd, Termux detection, sudo'd root install, and--uninstall --dry-runpreview. Runnable locally or via theinstall-e2eworkflow on demand.
How It's Built
Gormes is the artifact of an autonomous-porting methodology — a validation-gated planner → builder → TDD-slice loop that ports Hermes to Go one bounded vertical at a time. The runtime is the product; the methodology is the supporting evidence.
- Strategy: Gormes Success Plan
- Engineering blog: TrebuchetDynamics Engineering (RSS)
- Differentiator: v1.0 differentiator
- Toolkit extraction (
agentic-porting-kit): tracked as a Phase 8 row; until that public repo exists, the repo-local development skills are the inspectable placeholder.
Hermes-Agent, with upstream Git history preserved for attribution, remains the parity oracle.
Status
Latest public release: v0.2.12.
CI runs go test ./... -count=1, go run ./cmd/progress validate, and git diff --check. The single static binary ships for Linux, macOS, Windows, and Termux/Android. The current Linux build measures ~45.8 MB (benchmarks.json). WASI Whisper tiny.en runs at 3.78x realtime (benchmarks.json, 2026-05-10). Offline doctor peaks at ~25.6 MB RSS (benchmarks.json, 2026-05-16).
| Phase | Status | Shipped |
|---|---|---|
| Phase 1 — The Dashboard | ✅ | 6/6 subphases |
| Phase 2 — The Gateway | ✅ | 22/22 subphases |
| Phase 3 — The Black Box (Memory) | ✅ | 15/15 subphases |
| Phase 4 — The Brain Transplant | ✅ | 13/13 subphases |
| Phase 5 — The Final Purge | 🔨 | 21/23 subphases |
| Phase 6 — The Learning Loop (Soul) | 🔨 | 9/12 subphases |
| Phase 7 — Paused Channel Backlog | ✅ | 5/5 subphases |
| Phase 8 — Reputation & Publication | 🔨 | 3/7 subphases |
| Phase 9 — Design & Security Hardening | 🔨 | 6/7 subphases |
See CHANGELOG.md and SECURITY.md.
Built by Trebuchet Dynamics.
Hermes Agent lineage by Nous Research.
MIT license.
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found