ai-agent-board 0.2.8 → 0.2.10

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 +92 -80
  2. package/package.json +1 -1
package/README.md CHANGED
@@ -4,7 +4,7 @@
4
4
 
5
5
  **A visual dashboard for managing autonomous AI coding agents.**
6
6
 
7
- Kanban board • Real-time logs • Diff review • One-click PRs
7
+ Kanban board • Real-time chat • Diff review • One-click PRs
8
8
 
9
9
  [![npm version](https://img.shields.io/npm/v/ai-agent-board?color=blue&label=npm)](https://www.npmjs.com/package/ai-agent-board)
10
10
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
@@ -17,16 +17,18 @@ Kanban board • Real-time logs • Diff review • One-click PRs
17
17
 
18
18
  ## What is agent-board?
19
19
 
20
- `agent-board` is a local web dashboard that lets you **create tasks**, **assign them to AI coding agents**, and **monitor everything** from a Kanban-style interface — real-time execution logs, code diffs, and automatic PR creation on GitHub.
20
+ `agent-board` is a local web dashboard that lets you **create tasks**, **assign them to AI coding agents**, and **monitor everything** from a Kanban-style interface — real-time execution logs, live chat with the agent, code diffs, and automatic PR creation.
21
21
 
22
- It works with the coding CLIs you already have installed:
22
+ It auto-detects the coding CLIs you already have installed:
23
23
 
24
- | Agent | CLI Command | Status |
25
- |-------|-------------|--------|
26
- | **Claude Code** | `claude` | Supported |
27
- | **Codex** | `codex` | Supported |
28
- | **Copilot** | `copilot` | Supported |
29
- | **Gemini** | `gemini` | Supported |
24
+ | Agent | Command | Notes |
25
+ |-------|---------|-------|
26
+ | **Claude Code** | `claude` | Requires Anthropic API key or CLI login |
27
+ | **Codex** | `codex` | Requires OpenAI API key or CLI login |
28
+ | **Gemini** | `gemini` | Requires Google API key or CLI login |
29
+ | **OpenRouter** | _(API)_ | Use any model via OpenRouter API key |
30
+
31
+ > No CLI installed? Use **OpenRouter** to access 100+ models through the API without any CLI setup.
30
32
 
31
33
  ## Quick Start
32
34
 
@@ -36,14 +38,14 @@ npx ai-agent-board
36
38
 
37
39
  That's it. The dashboard opens automatically in your browser. No configuration files, no Docker, no database setup.
38
40
 
39
- > **Requirements:** Node.js >= 18. At least one supported AI coding CLI installed and authenticated.
41
+ > **Requirements:** Node.js >= 18. At least one supported AI coding CLI installed, or an OpenRouter API key.
40
42
 
41
43
  ### CLI Options
42
44
 
43
45
  ```
44
46
  npx ai-agent-board [options]
45
47
 
46
- --port <number> Server port (default: auto-detect)
48
+ --port <number> Server port (default: 51767)
47
49
  --no-open Don't open browser automatically
48
50
  --clear-cache Delete cached binary and re-download
49
51
  --version, -v Show version
@@ -54,37 +56,45 @@ npx ai-agent-board [options]
54
56
 
55
57
  ### Kanban Board
56
58
 
57
- Manage tasks across stages: **To Do** → **In Progress** → **In Review** → **Done**. Each task card shows its current status, assigned agent, and progress at a glance.
59
+ Manage tasks across stages: **To Do** &rarr; **In Progress** &rarr; **In Review** &rarr; **Done**. Each task card shows its current status, assigned agent, and progress at a glance.
58
60
 
59
- ### Two-Agent Workflow
61
+ ### Task Workflow
60
62
 
61
- Tasks go through a structured pipeline:
63
+ Describe what you need, and the agent handles the rest:
62
64
 
63
65
  ```
64
- You describe a task PM Agent generates You review & edit Dev Agent writes
65
- in natural language ───▶ a detailed spec ───▶ the spec ───▶ the code
66
- │
67
- ▼
68
- Review diff & merge
66
+ You describe a task Agent works in an You review the Approve & create
67
+ in natural language ───▶ isolated worktree ───▶ code changes ───▶ a Pull Request
69
68
  ```
70
69
 
71
- 1. **Create task** — describe what you need in plain language
72
- 2. **PM Agent** — generates a detailed technical specification
73
- 3. **Review spec** — edit, regenerate, or approve
74
- 4. **Dev Agent** — implements the approved specification
75
- 5. **Review & merge** — inspect the diff, then create a PR
70
+ 1. **Create task** &mdash; describe what you need in plain language
71
+ 2. **Agent codes** &mdash; the AI agent works in an isolated git worktree, with full context of your repo's stack and conventions
72
+ 3. **Chat in real-time** &mdash; send feedback, redirect the approach, or approve the agent's plan mid-execution
73
+ 4. **Review & merge** &mdash; inspect the diff, then create a PR directly from the dashboard
76
74
 
77
- ### Real-time Execution Logs
75
+ ### Real-time Chat
78
76
 
79
- Watch your AI agent work in real-time via Server-Sent Events (SSE). Every step, every file change, streamed directly to your browser.
77
+ Chat with the agent while it works. Send feedback, ask questions, or redirect its approach &mdash; all streamed live via SSE. The agent can present an implementation plan for your approval before writing any code.
80
78
 
81
79
  ### Diff Viewer
82
80
 
83
81
  Review all code changes before they go anywhere. Inline diff viewer shows exactly what the agent modified, added, or removed.
84
82
 
85
- ### GitHub Integration
83
+ ### Git Integration
84
+
85
+ - **GitHub** &mdash; create PRs, review comments, request changes, and track merge status
86
+ - **GitLab** &mdash; connect via Personal Access Token
87
+ - **Merge conflict resolution** &mdash; open VS Code directly at the worktree to resolve conflicts
88
+
89
+ ### Repository Management
90
+
91
+ - Scan your filesystem to discover local git repositories
92
+ - Auto-detect tech stack and conventions per repo
93
+ - Track learned patterns across tasks for smarter specs
94
+
95
+ ### Credential Management
86
96
 
87
- Create Pull Requests directly from the dashboard. Review PR comments, request changes, and track merge status — all without leaving the board.
97
+ API keys and tokens are stored locally with encryption. Connect your providers from the Settings page &mdash; no `.env` files needed.
88
98
 
89
99
  ### Dark Mode
90
100
 
@@ -94,51 +104,51 @@ Full light and dark theme support. Automatically follows your system preference.
94
104
 
95
105
  `agent-board` runs a lightweight local server that:
96
106
 
97
- 1. **Detects** which AI coding CLIs you have installed (Claude Code, Codex, Copilot, Gemini)
98
- 2. **Orchestrates** task execution by invoking the CLI agents in sandboxed worktrees
99
- 3. **Streams** real-time output to the browser via SSE
100
- 4. **Manages** code changes with git worktrees — your main branch stays untouched
101
- 5. **Creates** Pull Requests via the GitHub API when you approve changes
107
+ 1. **Detects** which AI coding CLIs you have installed (Claude Code, Codex, Gemini) or uses your OpenRouter API key
108
+ 2. **Assigns** a coding agent to work on your task in an isolated git worktree
109
+ 3. **Streams** real-time output, chat messages, and tool activity to the browser via SSE
110
+ 4. **Isolates** code changes in git worktrees &mdash; your main branch is never at risk
111
+ 5. **Creates** Pull Requests via the GitHub/GitLab API when you approve changes
102
112
 
103
- All data is stored in a local SQLite database. Nothing leaves your machine except the API calls your AI coding CLIs already make and the GitHub API calls for PR creation.
113
+ All data is stored in a local SQLite database. Nothing leaves your machine except the API calls your AI coding CLIs already make and the git provider API calls for PR creation.
104
114
 
105
115
  ## Task Lifecycle
106
116
 
107
117
  | Status | Description | Board Column |
108
118
  |--------|-------------|:------------:|
109
- | `draft` | Task created, awaiting spec generation | To Do |
110
- | `refining` | PM Agent generating specification | To Do |
111
- | `pending_approval` | Spec ready for your review | To Do |
112
- | `approved` | Spec approved, Dev Agent starting | In Progress |
113
- | `coding` | Dev Agent working | In Progress |
114
- | `awaiting_review` | Changes ready for your review | In Review |
115
- | `pr_created` | Pull Request created on GitHub | In Review |
116
- | `done` | PR merged | Done |
117
- | `failed` | Error during execution | — |
119
+ | `draft` | Task created, not yet started | To Do |
120
+ | `coding` | Agent working on implementation | In Progress |
121
+ | `plan_review` | Agent proposed a plan, waiting for your approval | In Progress |
122
+ | `review` | Changes ready for review / PR created | In Review |
123
+ | `merge_conflicts` | Merge conflicts detected, needs resolution | In Review |
124
+ | `changes_requested` | You requested changes on the PR | In Review |
125
+ | `done` | PR merged, task complete | Done |
126
+ | `canceled` | Canceled by user or PR closed | &mdash; |
127
+ | `failed` | Error during execution | &mdash; |
118
128
 
119
129
  ## Architecture
120
130
 
121
131
  ```
122
- ┌─────────────────────────────────────────────────────┐
123
- │ ai-agent-board │
124
- │ │
125
- │ ┌──────────────┐ ┌──────────────┐ ┌────────────┐ │
126
- │ │ Dashboard │ │ Server │ │ Shared │ │
127
- │ │ React SPA │◄─┤ Express API │ │ Types & │ │
128
- │ │ Vite 7 │ │ Bun runtime │ │ Schemas │ │
129
- │ │ TanStack │ │ SQLite │ │ Zod │ │
130
- │ │ Tailwind 4 │ │ SSE │ │ │ │
131
- │ └──────────────┘ └──────┬───────┘ └────────────┘ │
132
- │ │ │
133
- │ ┌──────┴───────┐ │
134
- │ │ CLI Runner │ │
135
- │ │ Orchestrator│ │
136
- │ └──────┬───────┘ │
137
- │ │ │
138
- │ ┌────────────┼────────────┐ │
139
- │ ▼ ▼ ▼ │
140
- │ claude codex gemini ... │
141
- └─────────────────────────────────────────────────────┘
132
+ ┌──────────────────────────────────────────────────────────┐
133
+ │ ai-agent-board │
134
+ │ │
135
+ │ ┌──────────────┐ ┌───────────────┐ ┌───────────────┐ │
136
+ │ │ Dashboard │ │ Server │ │ Shared │ │
137
+ │ │ React 19 │◄─┤ Express API │ │ Zod Schemas │ │
138
+ │ │ Vite 7 │ │ Bun runtime │ │ TypeScript │ │
139
+ │ │ TanStack │ │ SQLite WASM │ │ Types │ │
140
+ │ │ Tailwind 4 │ │ SSE + Chat │ │ │ │
141
+ │ └──────────────┘ └───────┬───────┘ └───────────────┘ │
142
+ │ │ │
143
+ │ ┌────────────┴────────────┐ │
144
+ │ │ Agent Orchestrator │ │
145
+ │ │ Task ──▶ AI Agent │ │
146
+ │ └────────────┬────────────┘ │
147
+ │ │ │
148
+ │ ┌───────────┬───┴───┬────────────┐ │
149
+ │ ▼ ▼ ▼ ▼ │
150
+ │ claude codex gemini OpenRouter │
151
+ └──────────────────────────────────────────────────────────┘
142
152
  ```
143
153
 
144
154
  ### Tech Stack
@@ -148,14 +158,14 @@ All data is stored in a local SQLite database. Nothing leaves your machine excep
148
158
  | **Frontend** | React 19, Vite 7, TypeScript, Tailwind CSS 4, shadcn/ui, TanStack Router & Query, Zustand |
149
159
  | **Backend** | Express, Bun, TypeScript, sql.js (SQLite via WASM), SSE |
150
160
  | **Shared** | Zod schemas, TypeScript types, npm workspaces |
151
- | **CLI** | Lightweight npx wrapper, platform-specific binary distribution |
161
+ | **CLI** | Lightweight npx wrapper, platform-specific binary distribution via Cloudflare R2 |
152
162
 
153
163
  ### Key Design Decisions
154
164
 
155
- - **SSE over WebSockets** — unidirectional streaming is sufficient for logs; simpler, works with any proxy/CDN, and browsers auto-reconnect
156
- - **SQLite (WASM)** — zero-config embedded database, no external server needed, cross-platform via WebAssembly
157
- - **Git worktrees** — agent works in isolated worktrees so your main branch is never at risk
158
- - **npx distribution** — single command to run, binary auto-downloaded and cached per platform
165
+ - **SSE over WebSockets** &mdash; unidirectional streaming is sufficient for logs and chat; simpler, works with any proxy, and browsers auto-reconnect
166
+ - **SQLite (WASM)** &mdash; zero-config embedded database, no external server needed, cross-platform via WebAssembly
167
+ - **Git worktrees** &mdash; agent works in isolated worktrees so your main branch is never at risk
168
+ - **npx distribution** &mdash; single command to run, binary auto-downloaded and cached per platform
159
169
 
160
170
  ## Development
161
171
 
@@ -168,33 +178,35 @@ npm install
168
178
  # Run in development mode (server + dashboard)
169
179
  npm run dev
170
180
 
171
- # Build all packages
181
+ # Build all packages (shared → server → dashboard)
172
182
  npm run build
173
183
 
174
184
  # Build specific packages
175
185
  npm run build:shared
176
186
  npm run build:server
177
187
  npm run build:dashboard
188
+
189
+ # Build standalone binaries
190
+ npm run build:binary:linux-x64
191
+ npm run build:binary:macos-x64
192
+ npm run build:binary:macos-arm64
193
+ npm run build:binary:win-x64
178
194
  ```
179
195
 
180
- The dashboard runs on **port 3003** (Vite dev server) and the server on **port 3000** (Bun). Vite proxies `/api` requests to the server automatically.
196
+ The dashboard dev server runs on **port 3003** and the backend on **port 51767**. Vite proxies `/api` requests to the server automatically.
181
197
 
182
198
  ### Environment Variables
183
199
 
184
- Create `packages/server/.env`:
200
+ Most configuration is handled through the **Settings** page in the dashboard. Environment variables are optional and mainly useful for development:
185
201
 
186
202
  ```env
187
- # Required: for AI agent capabilities
188
- OPENAI_API_KEY=sk-...
189
-
190
- # Optional: for GitHub PR creation
191
- GITHUB_TOKEN=ghp_...
192
-
193
- # Optional: server configuration
194
- PORT=3000
195
- DATABASE_PATH=./data/dash-agent.db
203
+ PORT=51767 # Server port (default: 51767)
204
+ DATABASE_PATH=./data/agent-board.db # SQLite database path
205
+ LOG_LEVEL=info # debug | info | warn | error
196
206
  ```
197
207
 
208
+ AI keys and git tokens are managed through the dashboard UI with encrypted local storage &mdash; no `.env` file needed for normal usage.
209
+
198
210
  ## Contributing
199
211
 
200
212
  Contributions are welcome. Please open an issue first to discuss what you'd like to change.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "ai-agent-board",
3
- "version": "0.2.8",
3
+ "version": "0.2.10",
4
4
  "type": "module",
5
5
  "description": "Dashboard for managing autonomous AI agent tasks — run with npx agent-board",
6
6
  "bin": {