hermes-launch 1.2.0 β†’ 2.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (4) hide show
  1. package/README.md +60 -48
  2. package/bin/hermes-launch.js +127 -583
  3. package/package.json +2 -2
  4. package/src/server.js +599 -0
package/README.md CHANGED
@@ -1,76 +1,88 @@
1
1
  # 🐚 Hermes Launch
2
2
 
3
- **One command. Zero to AI agent.**
3
+ **Web onboarding for Hermes Agent. Install, configure, and launch from your browser.**
4
4
 
5
5
  ```bash
6
6
  npx hermes-launch
7
7
  ```
8
8
 
9
- Hermes Launch installs, configures, and starts [Hermes Agent](https://github.com/NousResearch/hermes-agent) β€” the open-source AI agent framework by Nous Research. No faffing with Python venvs, curl pipes, or config files.
9
+ Opens a web wizard at `http://localhost:5050` that walks you through everything β€” no terminal fiddling, no config files, no curl pipes.
10
10
 
11
- ```bash
12
- # Full interactive setup (recommended for first use)
13
- npx hermes-launch
11
+ ## What it does
14
12
 
15
- # Non-interactive β€” smart defaults, go fast
16
- npx hermes-launch --quick
13
+ The wizard guides you through 5 steps:
17
14
 
18
- # Just start the web dashboard if Hermes is already installed
19
- npx hermes-launch --dashboard
20
- ```
15
+ 1. **Prerequisites** β€” Detects Python, curl, git. Installs missing ones automatically.
16
+ 2. **Install Hermes** β€” Downloads and installs Hermes Agent if not present.
17
+ 3. **Configure model** β€” Pick your LLM provider and model from a dropdown.
18
+ 4. **Enable tools** β€” Toggle which capabilities you want (web search, terminal, file ops, etc.).
19
+ 5. **Start dashboard** β€” Launches the Hermes web dashboard and opens it in your browser.
21
20
 
22
- ## What you get
21
+ Everything happens in the browser. The CLI just starts the server and gets out of the way.
23
22
 
24
- After running `npx hermes-launch`:
23
+ ## Usage
25
24
 
26
- - **Hermes Agent CLI** β€” `hermes` in your terminal. Chat, code, research, automate.
27
- - **Web dashboard** at `http://localhost:9119` β€” manage config, API keys, sessions.
28
- - **Tool sets** enabled out of the box: web search, terminal, file ops, skills, memory, vision, TTS, subagent delegation, cross-session search.
29
- - **Streaming** enabled β€” see responses as they generate.
30
- - **Zero config** β€” the setup wizard walks you through model/provider selection.
25
+ ```bash
26
+ # Start the setup wizard
27
+ npx hermes-launch
31
28
 
32
- ## After setup
29
+ # Custom port
30
+ npx hermes-launch --port 8080
33
31
 
34
- ```bash
35
- hermes # start a conversation
36
- hermes chat -q "analyze this repo" # one-shot
37
- hermes model # change provider/model
38
- hermes gateway setup # connect Telegram/Discord/Slack
39
- hermes tools list # see what tools are available
40
- hermes dashboard # open the web UI
41
- hermes doctor # health check
32
+ # Stop the background server
33
+ npx hermes-launch --stop
34
+
35
+ # Check if it's running
36
+ npx hermes-launch --status
42
37
  ```
43
38
 
44
- ## What's under the hood
39
+ ## How it works
45
40
 
46
- Hermes Launch is a thin wrapper β€” it:
41
+ ```
42
+ npx hermes-launch
43
+ β”‚
44
+ β–Ό
45
+ Starts Node.js HTTP server (port 5050) ──detached──▢ runs forever
46
+ β”‚ β”‚
47
+ β–Ό β”‚
48
+ Opens browser to http://localhost:5050 β”‚
49
+ β”‚ β”‚
50
+ β–Ό β”‚
51
+ Web wizard guides through: β”‚
52
+ 1. Prerequisites check β”‚
53
+ 2. Hermes install (if needed) β”‚
54
+ 3. Model/provider selection β”‚
55
+ 4. Tool enablement β”‚
56
+ 5. Dashboard start β”‚
57
+ β”‚ β”‚
58
+ β–Ό β”‚
59
+ Done β†’ link to Hermes dashboard at localhost:9119 β”‚
60
+ β”‚
61
+ npx hermes-launch --stop ───────────────────────▢ kills server
62
+ ```
63
+
64
+ - **Zero npm dependencies** β€” built-in Node.js `http` module
65
+ - **Cross-platform** β€” macOS, Linux, Windows (with auto-install of missing deps)
66
+ - **Detached server** β€” runs in background after setup, survives terminal close
67
+ - **No pipe-to-bash** β€” safe for enterprise environments
47
68
 
48
- 1. Checks prerequisites (Python 3.10+, curl, git)
49
- 2. Installs Hermes Agent via the official install script if missing
50
- 3. Guides you through `hermes setup` to pick a model provider
51
- 4. Enables the core toolsets
52
- 5. Starts `hermes dashboard --tui` on port 9119
53
- 6. Prints next steps
69
+ ## What's different from v1
54
70
 
55
- Everything lives in `~/.hermes/` β€” config, sessions, logs, skills.
71
+ | Before (v1) | After (v2) |
72
+ |---|---|
73
+ | CLI does everything in terminal | Web wizard does everything in browser |
74
+ | Tries to open dashboard directly | Opens onboarding page first |
75
+ | Dashboard dies when script exits | Server is detached, runs forever |
76
+ | No visual feedback | Step-by-step UI with status |
77
+ | Manual tool config | Toggle to enable/disable tools |
78
+ | Browser opening races with server | Waits for server to be ready |
56
79
 
57
80
  ## Requirements
58
81
 
59
82
  - **Node.js** 18+ (comes with npx)
60
- - **Python** 3.10+ (system or homebrew)
83
+ - **Python** 3.10+ (auto-installed via winget/choco/brew on some platforms)
61
84
  - **curl**, **git**
62
- - macOS, Linux, or WSL
63
-
64
- ## Why not just curl the install script?
65
-
66
- You can! `curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash` works great. Hermes Launch adds:
67
-
68
- - Prerequisite checking with clear error messages
69
- - Interactive setup flow with prompts
70
- - Smart defaults for tool enablement
71
- - Dashboard auto-start
72
- - A clean post-install summary
73
- - No pipe-to-bash pattern (safer for enterprise environments)
85
+ - macOS, Linux, or Windows
74
86
 
75
87
  ## License
76
88