Open Source — v1.5.15

Orchestrate your
AI dev squad.

Run multiple Claude agents with dependency graphs, wave execution, and live observability. Build features in parallel — from a single command center.

E-Commerce Platform
Graph Timeline Live Board
Wave 0 Wave 1 Wave 2 Auth API opus · 4m 12s · done Product Catalog sonnet · running DB Schema haiku · 1m 02s · done Search Service sonnet · queued Payment opus · waiting deps Admin Dashboard sonnet · waiting deps E2E Tests opus · wave 2 Deploy Script haiku · wave 2
Product Catalog sonnet
Tokens24.3k
Cost$0.08
Tools18
Time2m 34s
Edit src/models/product.js
Auth API opus
Tokens48.1k
Cost$0.24
Tools31
Time4m 12s
Completed · 7 files
DB Schema haiku
Tokens8.2k
Cost$0.02
Tools9
Time1m 02s
Completed · 3 files
DAG Dependencies
Wave Execution
Live Terminal
Auto-PR Creation
Session Replay
Conflict Detection
Auto-Test
Exponential Backoff
Webhooks
Planning Mode
Token Analytics
Git Automation
DAG Dependencies
Wave Execution
Live Terminal
Auto-PR Creation
Session Replay
Conflict Detection
Auto-Test
Exponential Backoff
Webhooks
Planning Mode
Token Analytics
Git Automation

Three views. Total control.

Monitor, debug, and manage your entire AI pipeline from the orchestration command center.

Auth
API
Payment
Tests
Deploy
🖨

Dependency Graph

Interactive DAG with draggable nodes, conditional edges, and wave grouping. Shift+drag to create dependencies.

Auth API
Catalog
Schema
Payment
Tests
📅

Execution Timeline

Gantt-style view with real-time progress, NOW marker, wave groups, and hover tooltips with cost breakdown.

14:23:01 Catalog Edit src/models/product.js
14:23:00 Auth Bash npm test ✓ 12 passed
14:22:58 Catalog Read src/routes/products.js
14:22:55 Auth Write src/middleware/jwt.js +42
14:22:52 ⚠ CONFLICT config.js — 2 agents
📡

Live Observability

Real-time feed of all tool calls, file heatmap with conflict detection, and per-agent activity cards.

From idea to production.
In three moves.

1

Define your tasks

Describe features, bugs, or refactors. Pick models, set priorities, and define dependencies between tasks with drag-and-drop.

2

Agents orchestrate

Tasks execute in dependency-aware waves. Each agent creates a branch, writes code, runs tests, and commits — all visible in the live terminal.

3

Review & ship

Review diffs, auto-test results, and token costs. Approve to auto-create PRs, or request changes — Claude iterates on your feedback.

Task Description
"Build a payment integration with Stripe. Handle subscriptions, webhooks, and refunds."
Model
⬤ Opus
Type
Feature
Priority
High
Dependencies
✓ Auth API ✓ DB Schema
Attached Context
📄 stripe-api-docs.md ⚙ coding-standards
▶ Wave 0 — 3 agents parallel
Read src/auth/middleware.js
Write src/auth/jwt.service.js +94 lines
Bash npm test ✓ 18 passed
Bash git commit -m "feat: JWT auth service"
✓ Auth API completed • 4m 12s • $0.24
▶ Wave 1 — dependencies resolved, starting 3 agents
⌛ Payment Integration waiting for Auth API
Read src/services/stripe.js
Write src/payments/checkout.js +158 lines
Write src/payments/webhooks.js +82 lines
12Files
+847Added
-23Removed
$0.34Cost
src/payments/checkout.js
+ import Stripe from 'stripe';
+ import { validateSession } from '../auth/jwt.service';
 
+ export async function createCheckoutSession(req, res) {
+ const session = await validateSession(req.headers.auth);
+ const checkout = await stripe.checkout.sessions.create({
+ customer: session.customerId,
+ line_items: req.body.items,
+ });
✓ Approve & Create PR
↻ Request Changes

Everything you need to
run an AI dev squad.

From orchestration to deployment, Claude Board handles the entire lifecycle.

Live Terminal

Watch each agent's tool calls, file edits, and bash commands in real-time with split-terminal support.

🔌

Git Automation

Auto-create branches per task. One-click PR creation on completion. Full commit history tracking.

🔄

Auto-Retry & Backoff

Failed tasks retry automatically with exponential backoff. Configurable max retries with 30s×2^n delay.

📊

Usage Analytics

Token tracking, cost analysis per agent, model breakdown, 30-day trends, and rate limit monitoring.

🚩

Auto-Test & Review

Auto-run test suites on task completion. Failed tests trigger auto-revision. Diff viewer with inline comments.

🔔

Webhooks

Slack, Discord, Teams, or custom endpoints. Event filtering per project with one-click test delivery.

📄

Context & Templates

Snippets inject rules into every prompt. Templates with variable substitution. File attachments for context.

🛠

Claude Manager

Visual dashboard for MCP servers, plugins, agents, and permissions. One place for your Claude environment.

🔄

Session Replay

Replay any agent's entire session step-by-step. Debug issues, understand decisions, learn from outputs.

🔒

Planning Mode

Put your board in planning mode. Arrange tasks, set dependencies, and configure waves before execution.

See it in action.

Click to explore different views of Claude Board.

My Project
Board List Timeline Graph
Backlog 3
feature
Dark mode support
sonnet · #CB-12
bugfix
Fix login redirect
haiku · #CB-13
refactor
Clean up utils
haiku · #CB-14
Running 2
refactor
Migrate API to v2
Running
feature
User auth system
Running
Testing 1
feature
Profile settings
opus · 3m · #CB-10
Done 2
docs
Update API docs
2m · #CB-8
test
Integration tests
5m · #CB-9
Failed 1
bugfix
Cache invalidation
3/3 retries · #CB-7
Dashboard
9
Total Tasks
2
Running
5
Completed
82.4k
Tokens Used
Token Usage (7 days)
MonTueWedThuFriSatSun
By Model
Opus42%
Sonnet38%
Haiku20%
Live Terminal Agent Running
Read src/routes/api.js
Read src/models/user.js
Write src/routes/auth.js +127 lines
Write src/middleware/jwt.js +84 lines
Edit src/app.js +3 -1
Bash npm test
✓ 23 tests passed (0 failed)
Bash git add -A && git commit -m "feat: JWT auth middleware"
[feature/CB-101] 1 file changed, 211 insertions(+), 1 deletion(-)
Read src/services/stripe.js
Orchestration · Graph
GraphTimelineLive
Wave 0Wave 1Wave 2
Auth API4m
Catalog2m
Schema1m
Search
Payment
Admin
E2E Tests
Deploy
Active Agents
Catalogsonnet
24.3k tokens · 18 tools
Auth APIopus
✓ Completed · 7 files
Schemahaiku
✓ Completed · 3 files
Planning ModePLANNING
1Set up database schemahaikuWave 0
2Build authentication APIopusWave 0
3Product catalog servicesonnetWave 0
4Payment integrationdepends on #1, #2opusWave 1
5Search servicedepends on #3sonnetWave 1
6E2E test suitedepends on #4, #5opusWave 2
▶ Start Execution
+ Add Task
AnalyticsLast 30 days
156
Tasks Run
89%
Success Rate
1.2M
Tokens
3.2m
Avg Duration
Tasks by Status
Done139
Running4
Failed13
Backlog0
Cost Breakdown
Opus$12.40
Sonnet$4.80
Haiku$0.90
Activity Log
14:23completedAuth API· 7 files · 4m 12s · opus
14:22testingProfile Settings· auto-test running
14:20startedProduct Catalog· branch created
14:18completedDB Schema· 3 files · 1m 02s · haiku
14:15failedCache Layer· retry 3/3 exhausted
14:12pr-createdAPI Docs· #47 opened
14:10retryCache Layer· attempt 3 · backoff 2m
14:05completedIntegration Tests· 5 files · 5m 30s
New Task
Description
"Add Stripe payment webhooks with signature verification"
Model
Opus
Type
Feature
Priority
High
Dependencies
✓ Auth API✓ DB Schema
Create Task

Get Claude Board.

Free & open source. Requires Claude Code CLI.

Or build from source: git clone && npm install && npx tauri dev

Ready to orchestrate your AI squad?

Dependency graphs. Wave execution. Live observability. All running on your machine.

Star on GitHub Read Docs