open-claude

Quick Start • Dashboard • Docs • Changelog • Contributing • MIT License

Disclaimer: OpenClaude is an independent, unofficial open-source project. It is not affiliated with, endorsed by, or sponsored by Anthropic. "Claude" and "Claude Code" are trademarks of Anthropic, PBC. This project integrates with Claude Code as a third-party tool and requires users to provide their own installation and credentials.

What It Is

OpenClaude is an open source, unofficial toolkit compatible with Claude Code and other LLM tooling. It is designed to integrate with Claude Code capabilities: native agents, skills, slash commands, MCP integrations, and the Claude CLI.

It turns a single Claude Code installation into a team of 9 specialized agents — each with its own domain, skills, persistent memory, and automated routines. The result is a production system that runs daily operations for a founder/CEO: from morning briefings to financial reports, community monitoring, social analytics, and end-of-day consolidation.

This is not a chatbot. It is a real operating layer that runs routines, generates HTML reports, syncs meetings, triages emails, monitors community health, tracks financial metrics, and consolidates everything into a unified dashboard — all automated.

Why Claude Code?

• Native agent system — agents are .md files with system prompts, not code
• Skills as instructions — teach Claude new capabilities via markdown, not plugins
• MCP integrations — first-class support for Google Calendar, Gmail, GitHub, Linear, Telegram, and more
• Slash commands — /clawdia, /flux, /pulse invoke agents directly
• Persistent memory — CLAUDE.md + per-agent memory survives across sessions
• CLI-first — runs anywhere Claude Code runs (terminal, VS Code, JetBrains, web)

Key Features

• 10 Core Agents + Custom — Ops, Finance, Projects, Community, Social, Strategy, Sales, Courses, Personal, Knowledge — plus user-created custom-* agents (gitignored)
• ~68 Skills — organized by domain prefix (social-, fin-, int-, prod-, mkt-, gog-, obs-, discord-, pulse-, sage-)
• 7 Core + 20 Custom Routines — daily, weekly, and monthly ADWs managed by a scheduler (core routines ship with the repo; custom routines are user-created and gitignored)
• Web Dashboard — React + Flask app with auth, roles, web terminal, service management
• 17 Integrations — Google Calendar, Gmail, Linear, GitHub, Discord, Telegram, Stripe, Omie, Fathom, Todoist, YouTube, Instagram, LinkedIn, Evolution API, Evolution Go, Evo CRM, and more
• 2 core + custom HTML report templates — dark-themed dashboards for every domain
• Persistent Memory — two-tier system (CLAUDE.md + memory/) with LLM Wiki pattern: ingest propagation, weekly lint, centralized index, and operation log
• Knowledge Base — optional semantic search via MemPalace (local ChromaDB vectors, one-click install)
• Full Observability — JSONL logs, execution metrics, cost tracking per routine

Screenshots

Overview	Chat
Integrations	Costs

Integrations

Connect your existing tools via MCP servers, API clients, or OAuth:

Social media accounts (YouTube, Instagram, LinkedIn) are connected via OAuth through the dashboard.

Prerequisites

The setup wizard (make setup) checks for all prerequisites before proceeding.

Quick Start

Method 1 — One command (recommended)

npx @evoapi/open-claude

This downloads and runs the interactive setup wizard automatically.

Method 2 — Manual clone

git clone https://github.com/EvolutionAPI/open-claude.git
cd open-claude

# Interactive setup wizard — checks prerequisites, creates config files
make setup

What the wizard does

The wizard:

• Checks that Claude Code, uv, Node.js are installed
• Asks for your name, company, timezone, language
• Lets you pick which agents and integrations to enable
• Generates config/workspace.yaml, .env, CLAUDE.md, and workspace folders
• Builds the dashboard frontend

2. Configure API keys

nano .env

Add keys for the integrations you enabled. Common ones:

# Discord (for community monitoring)
DISCORD_BOT_TOKEN=your_token
DISCORD_GUILD_ID=your_guild_id

# Stripe (for financial routines)
STRIPE_SECRET_KEY=sk_live_...

# Telegram (for notifications)
TELEGRAM_BOT_TOKEN=your_token
TELEGRAM_CHAT_ID=your_chat_id

# Social media (connect via dashboard Integrations page)
SOCIAL_YOUTUBE_1_API_KEY=...
SOCIAL_INSTAGRAM_1_ACCESS_TOKEN=...

See .env.example for all available variables.

3. Start the dashboard

make dashboard-app

Open http://localhost:8080 — the first run shows a setup wizard where you:

• Configure your workspace (name, company, agents, integrations)
• Create your admin account
• License is activated automatically in the background

4. Start automated routines

make scheduler

Runs all enabled routines on schedule (morning briefing, email triage, community pulse, financial reports, etc). Configure schedules in config/routines.yaml.

5. Use Claude Code

Open Claude Code in the project directory — it reads CLAUDE.md automatically.

# Use slash commands to invoke agents
/clawdia       # Ops — agenda, emails, tasks, decisions
/flux          # Finance — Stripe, ERP, cash flow, reports
/atlas         # Projects — Linear, GitHub, sprints, milestones
/pulse         # Community — Discord, WhatsApp, sentiment, FAQ
/pixel         # Social media — content, calendar, analytics
/sage          # Strategy — OKRs, roadmap, competitive analysis
/nex           # Sales — pipeline, proposals, qualification
/mentor        # Courses — learning paths, modules
/kai           # Personal — health, habits, routine
/oracle        # Knowledge — workspace docs, how-to, configuration

# Or just describe what you need — Claude routes to the right agent

Web Dashboard

A full web UI at http://localhost:8080:

make dashboard-app   # Start Flask + React on :8080

Architecture

User (human)
    |
    v
Claude Code (orchestrator)
    |
    +-- Clawdia   — ops: agenda, emails, tasks, decisions, dashboard
    +-- Flux      — finance: Stripe, ERP, MRR, cash flow, monthly close
    +-- Atlas     — projects: Linear, GitHub, milestones, sprints
    +-- Pulse     — community: Discord, WhatsApp, sentiment, FAQ
    +-- Pixel     — social: content, calendar, cross-platform analytics
    +-- Sage      — strategy: OKRs, roadmap, prioritization, scenarios
    +-- Nex       — sales: pipeline, proposals, qualification
    +-- Mentor    — courses: learning paths, modules
    +-- Kai       — personal: health, habits, routine (isolated domain)

Each agent has:

• System prompt in .claude/agents/
• Slash command in .claude/commands/
• Persistent memory in .claude/agent-memory/
• Related skills in .claude/skills/

Workspace Structure

open-claude/
├── .claude/
│   ├── agents/          — 9 agent system prompts
│   ├── commands/        — 9 slash commands
│   ├── skills/          — ~68 skills by prefix
│   └── templates/html/  — 2 core + custom HTML templates
├── ADWs/
│   ├── runner.py        — execution engine (logs + metrics + notifications)
│   ├── routines/         — 7 core routine scripts (shipped with repo)
│   └── routines/custom/  — 20 custom routines (user-created, gitignored)
├── dashboard/
│   ├── backend/         — Flask + SQLAlchemy + WebSocket
│   └── frontend/        — React + TypeScript + Tailwind
├── social-auth/         — OAuth multi-account app
├── config/              — workspace.yaml, routines.yaml
├── workspace/           — user data folders (gitignored content)
├── setup.py             — CLI setup wizard
├── scheduler.py         — automated routine scheduler
├── Makefile             — 44+ make targets
└── CLAUDE.template.md   — template for generated CLAUDE.md

Workspace folders (workspace/daily-logs/, workspace/projects/, etc.) are created by setup — content is gitignored, only structure is tracked.

Commands

# Setup & Dashboard
make setup           # Interactive setup wizard
make dashboard-app   # Start web dashboard on :8080


# Routines
make scheduler       # Start automated routine scheduler
make morning         # Run morning briefing
make triage          # Run email triage
make community       # Run community pulse
make fin-pulse       # Run financial pulse
make eod             # Run end-of-day consolidation
make memory-lint     # Run memory health check (contradictions, stale data, gaps)
make weekly          # Run weekly review

# Observability
make logs            # Show latest JSONL log entries
make metrics         # Show per-routine metrics (runs, cost, tokens)
make help            # List all available commands

Documentation

License

MIT License. See LICENSE for details.

An unofficial community toolkit for Claude Code
_{Built by Evolution Foundation — not affiliated with Anthropic}