imcodes 2026.3.241 → 2026.3.243

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 (2) hide show
  1. package/README.md +34 -19
  2. package/package.json +1 -1
package/README.md CHANGED
@@ -1,6 +1,6 @@
1
1
  # IM.codes
2
2
 
3
- A chat interface built for talking to AI coding agents. Not Slack, not Discord, not Telegram — something actually designed for the job.
3
+ A purpose-built interface for managing AI coding agents. Remote terminal, multi-agent orchestration, and session management — from any device.
4
4
 
5
5
  ## Screenshots
6
6
 
@@ -25,24 +25,37 @@ A chat interface built for talking to AI coding agents. Not Slack, not Discord,
25
25
 
26
26
  ## Why
27
27
 
28
- Every existing chat tool is built for humans talking to humans. When you're working with Claude Code, Codex, or Gemini CLI, you need terminal output, diff views, session management, multi-agent coordination — none of which fit into a 4096-char message box with emoji reactions.
28
+ AI coding agents live in terminals. Managing them means SSH sessions, split panes, copy-pasting between windows. It doesn't scale — not across machines, not from a phone, and definitely not when you're running multiple agents on a single task.
29
29
 
30
- I wanted a tool I could fully customize. Something that speaks the language of code, not social media. So I built one.
30
+ IM.codes gives you a single control plane: remote terminals that work like SSH without the SSH, a chat layer that understands agent output, and multi-agent workflows that let you pit different models against each other.
31
31
 
32
32
  This is a personal project. I haven't written any code myself — it was built almost entirely by [Claude Code](https://github.com/anthropics/claude-code), with significant contributions from [Codex](https://github.com/openai/codex) and [Gemini CLI](https://github.com/google-gemini/gemini-cli).
33
33
 
34
- ## What it does
34
+ ## Features
35
35
 
36
- - **Remote agent control** — Talk to your agents from a browser or phone. No SSH, no VPN.
37
- - **Terminal + Chat modes** — Switch between raw terminal (the native CLI experience) and a structured chat view.
38
- - **Multi-server, multi-session** — Manage agents across machines from one dashboard.
39
- - **Real-time streaming** — Live terminal output, no message limits, no rate throttling.
40
- - **Sub-sessions** — Spawn additional agents from within a session. Run parallel tasks with full visibility.
41
- - **P2P multi-agent discussions** — Route a topic through multiple agents in sequence. Each agent reads prior contributions and adds their own. Supports discuss, audit, review, and brainstorm modes. Works across Claude Code, Codex, and Gemini CLI — including sandboxed agents.
42
- - **File upload & download** — Upload files from browser or phone (including camera). Files are relayed to the daemon and referenced in chat with `@path`. Download project files directly from the built-in file browser.
43
- - **Push notifications** — Get notified on your phone when an agent needs attention.
44
- - **OTA updates** — Daemon self-upgrades via npm. Trigger from the web UI for one device or all devices at once.
45
- - **Fully customizable** — It's your UI. Add whatever you need — diff viewers, approval flows, custom scripts.
36
+ ### Remote Terminal
37
+
38
+ Full terminal access to your agent sessions from any browser — no SSH, no VPN, no port forwarding. Switch between raw terminal mode (the native CLI experience) and a structured chat view with parsed tool calls, thinking blocks, and streaming output. Real-time PTY streaming with zero message limits.
39
+
40
+ ### Multi-Agent Discussions & Audit
41
+
42
+ Single-model output shouldn't be trusted blindly. Spawn quick discussion rounds where multiple agents — across different providers — review, audit, or brainstorm on the same topic. Each agent reads prior contributions and adds their own. Modes include `discuss`, `audit`, `review`, and `brainstorm`. Works across Claude Code, Codex, and Gemini CLI, including sandboxed agents.
43
+
44
+ ### Multi-Server, Multi-Session Management
45
+
46
+ Connect multiple dev machines to one dashboard. Each machine runs a lightweight daemon that manages local agent sessions via tmux. See all servers and sessions at a glance — start, stop, restart, or switch between them instantly. Sub-sessions let you spawn additional agents from within a running session for parallel tasks.
47
+
48
+ ### File Browser & Preview
49
+
50
+ Browse project files, preview any file with syntax highlighting, and view git diffs with inline change comparison — all from the browser. Upload files from browser or phone (including camera capture), referenced in chat with `@path`.
51
+
52
+ ### Mobile & Notifications
53
+
54
+ Full mobile support with biometric auth. Push notifications when an agent finishes a task or needs attention — so you don't have to watch a spinner.
55
+
56
+ ### OTA Updates
57
+
58
+ Daemon self-upgrades via npm. Trigger from the web UI for one device or all devices at once.
46
59
 
47
60
  ## Architecture
48
61
 
@@ -74,7 +87,7 @@ imcodes bind https://app.im.codes/bind/<api-key>
74
87
 
75
88
  This binds your machine, starts the daemon, and registers it as a system service.
76
89
 
77
- ## Self-Host (One-Click Deploy)
90
+ ## Self-Host
78
91
 
79
92
  Deploy your own IM.codes server with Docker Compose. Includes automatic HTTPS via Caddy and auto-updates via Watchtower.
80
93
 
@@ -84,19 +97,21 @@ git clone https://github.com/im4codes/imcodes.git
84
97
  cd imcodes
85
98
  cp .env.example .env
86
99
 
87
- # 2. Edit .env — set your domain and generate secrets
100
+ # 2. Edit .env — set your domain, admin password, and secrets
88
101
  # DOMAIN=im.example.com
102
+ # DEFAULT_ADMIN_PASSWORD=your-secure-password
89
103
  # POSTGRES_PASSWORD=$(openssl rand -hex 16)
90
104
  # JWT_SIGNING_KEY=$(openssl rand -hex 32)
91
- # BOT_ENCRYPTION_KEY=$(openssl rand -hex 32)
92
105
 
93
106
  # 3. Start
94
107
  docker compose up -d
95
108
  ```
96
109
 
97
- That's it. Caddy automatically provisions a Let's Encrypt certificate for your domain. Watchtower checks for new images every 5 minutes and updates automatically.
110
+ Caddy automatically provisions a Let's Encrypt certificate for your domain. Watchtower checks for new images every 5 minutes and updates automatically.
111
+
112
+ Point your DNS A record to the server IP, then visit `https://your-domain`.
98
113
 
99
- Point your DNS A record to the server IP, then visit `https://your-domain`. Default login: `admin` / password shown in `docker compose logs server` on first run.
114
+ Login with `admin` and the password you set in `DEFAULT_ADMIN_PASSWORD`. If not set, a random password is generated on first run — check `docker compose logs server` to find it.
100
115
 
101
116
  After logging in, bind your dev machine:
102
117
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "imcodes",
3
- "version": "2026.3.241",
3
+ "version": "2026.3.243",
4
4
  "description": "A chat interface for AI coding agents — control Claude Code, Codex, Gemini CLI, OpenCode from browser and mobile",
5
5
  "license": "GPL-3.0-only",
6
6
  "repository": {