Claw-Company / clawcompany
Build for OPC(One Person Company), Every human being is a chairman. Multiple roles, multiple agents, multiple suppliers. Your Claws company, one key to run them all.
View on GitHubAI Architecture Analysis
This repository is indexed by RepoMind. By analyzing Claw-Company/clawcompany in our AI interface, you can instantly generate complete architecture diagrams, visualize control flows, and perform automated security audits across the entire codebase.
Our Agentic Context Augmented Generation (Agentic CAG) engine loads full source files into context on-demand, avoiding the fragmentation of traditional RAG systems. Ask questions about the architecture, dependencies, or specific features to see it in action.
Repository Overview (README excerpt)
Crawler viewClawCompany > **Build for OPC (One Person Company). Every human being is a chairman.** > > Multiple roles, multiple agents, multiple suppliers. > Your Claws company, one key to run them all. ClawCompany is the infrastructure for OPC — One Person Companies powered by AI. You are the Chairman. Your AI team (CEO, CTO, CFO, CMO, Engineers, Researchers, Analysts) executes autonomously. You set the goal — they figure out the rest. **Website** · **Docs** · **Example Mission** · **ClawAPI** --- The OPC model You don't need to hire anyone. You don't need to manage anyone. You give your company a mission, and it runs. **Total: 3 minutes, $0.12.** Your AI company just delivered a professional report. If you ran the same mission using only Opus? **$3.50.** ClawCompany's multi-model architecture is **30x cheaper**. → See the full example mission with cost breakdown --- Quick start — 3 steps, 30 seconds That's it. The wizard guides you through everything: Now give it a mission: No config files. No JSON to edit. No Docker. No proxy. No server to start. Just answer 3 questions. > **Requirements:** Node.js 20+. Get a ClawAPI key at clawapi.org. --- Your AI team One ClawAPI key activates your entire company — 9 roles across 4 models: | Role | Model | Cost (in/out per 1M) | What they do | |------|-------|---------------------|--------------| | **CEO** | | $5 / $25 | Decomposes missions, coordinates departments, quality gate | | **CTO** | | $2.50 / $15 | Technical architecture, code review, system design | | **CFO** | | $0.25 / $2 | Financial analysis, budgets, projections (reasoning model) | | **CMO** | | $3 / $15 | Marketing strategy, content creation, brand voice | | **Researcher** | | $3 / $15 | Deep research, competitive analysis | | **Analyst** | | $0.25 / $2 | Data analysis, pattern detection (reasoning model) | | **Engineer** | | $2.50 / $15 | Code implementation, debugging, testing | | **Secretary** | | $0.25 / $1.50 | Briefings, summaries, formatting | | **Worker** | | $0.25 / $1.50 | Data collection, routine tasks | **Every role is fully customizable.** Rename, swap models, change providers, add your own roles. --- Why multi-model? The right model for the right job Professional work deserves professional models. Routine work deserves fast, cheap models. ClawCompany assigns automatically — faster execution, better results, 27x lower cost. | Step | Role | Model | Cost | If all Opus | |------|------|-------|------|-------------| | Decompose mission | CEO | opus | $0.027 | $0.027 | | Data collection | Worker | flash-lite | $0.001 | ~$0.40 | | Deep analysis | Researcher | sonnet | $0.026 | ~$0.45 | | Market positioning | CMO | sonnet | $0.009 | ~$0.45 | | Report formatting | Secretary | flash-lite | $0.001 | ~$0.40 | | **Total** | | | **$0.12** | **$3.50** | **96% cheaper. 27x less cost.** Routine work goes to flash-lite at $0.001 per call instead of Opus at $0.40+. The quality is identical because these tasks don't need deep reasoning. At scale: | Missions/day | ClawCompany | All-Opus | Monthly savings | |---|---|---|---| | 10 | $0.64/day | $17.30/day | **$499/month** | | 100 | $6.40/day | $173/day | **$4,998/month** | | 1,000 | $64/day | $1,730/day | **$49,980/month** | --- How it works --- Open model supply layer — 5 built-in providers ClawCompany is **not** locked to ClawAPI. Five providers are built in, and you can add more: | # | Provider | What you get | |---|----------|-------------| | 1 | **ClawAPI** ★ | 1 key, 8 models, crypto-native. The default. | | 2 | **Anthropic** | Claude models direct | | 3 | **OpenAI** | GPT models direct | | 4 | **Google Gemini** | Gemini models direct | | 5 | **Ollama** (local) | Run models locally, free | ★ = default supplier, always first Need another provider? Add it yourself, no approval needed: --- 💻 Code Manager — ∞ terminals. One Dashboard. Run Claude Code, Codex, Cursor, and any CLI tool from browser tabs. Full PTY terminal with ANSI colors and interactive input. • 7 projects. 5 running. 2 stopped. One glance. • Start all → green dots. Stop all → done. • Process completes → phone buzzes (Telegram/Discord). • Real terminal. Not a text viewer. --- 🚀 Company Templates — Switch your org chart in one click Pre-built role configurations for different business types. Switch instantly, keep your custom roles. • 🦞 **Default** — 9 roles, 4 AI models, general purpose • 🚀 **YC Startup** — 7 roles, YC methodology, ship fast • More coming: Media Studio, Trading Desk, E-Commerce --- ClawMarket — download a company, one click, it runs Browse templates, skills, and tools. Install and run instantly. | Category | Examples | Status | |----------|---------|--------| | **Templates** | Trading Desk, Content Agency, Dev Shop, Legal Firm, E-commerce, Consulting, Solo Founder | 8 available | | **Skills** | Web Scraper, PDF Generator, Crypto Price Feed, Email Sender, GitHub Operations | 5 available | | **Tools** | Code execution, browser automation | Coming soon | | **Services** | Slack, HubSpot, Stripe integrations | Coming soon | One click = working company. No configuration needed. --- Agent tools — agents that do things, not just say things 6 built-in tools. Your agents can search the web, read pages, call APIs, write files, and execute code — autonomously. | Tool | What it does | |------|-------------| | **Web Search** | Search the web for real-time info (DuckDuckGo, no API key needed) | | **Web Fetch** | Read any web page, strip HTML, return clean text | | **HTTP Requests** | Call any API — GET, POST, PUT, DELETE | | **Filesystem** | Read, write, list, delete files | | **Shell** | Execute shell commands (30s timeout) | | **Code Interpreter** | Run JavaScript or Python code | | **Browser Use** | Navigate pages, click, type, screenshot, run JS | Agents use a **think → act → observe** loop: reason about the task, use tools, observe results, repeat until done. Verified: Researcher made 16 tool calls in one mission to gather rea…