AI AGENT FRAMEWORK FOR CLAUDE CODE
You can build the app. Elven builds the business around it — strategy, marketing, finance, compliance, operations, and the 32 specialized agents that make them work together inside Claude Code.
Free tier available • One-time purchase • Works with your existing CLAUDE.md • 14-day refund
New to Claude Code? Elven is the fastest way to see what's possible.
Already shipping code? Elven is the business team you haven't hired yet.
What You Need to Run Elven
✓ Claude Code CLI or any agentic IDE (Cursor, Windsurf, etc.)
✓ Claude API subscription (~$20/month for typical usage)
✓ Your project codebase (new or existing)
THE PROCESS
From Idea to Revenue
Most AI tools help you do tasks. Elven gives you a methodology — a structured path from raw idea to shipped product, with specialized agents at every stage.
WHO IT'S FOR
Building a product is 20% code. Elven covers the other 80%.
32 agents. 22 of them aren't engineers. Here's what that means for you.
I can build anything. I can’t sell anything.
Your app works. It might have been working for weeks. But it’s 2am and you’re googling “how to write a privacy policy” instead of shipping features. The gap between “it works” and “someone pays for it” isn’t technical — it’s everything else. Positioning. Pricing. Marketing. Content. Go-to-market.
Elven gives you 22 non-engineering agents: a strategist who validates your positioning (Arthur), a marketer who plans your launch (Hermes), a finance agent who models your pricing (Moneta), a content strategist who writes your copy (Brigid), and a compliance agent who handles governance (Censoria).
I can design the experience. I can’t launch the business.
You can prototype, build, and ship a beautiful product. But somewhere between “looks great” and “people are paying” there’s a gap nobody warned you about. Business strategy. Pricing models. Compliance. SEO. Content that ranks. The product isn’t the hard part. Turning it into a business is.
Elven gives you the full stack — engineering agents when you need code support, plus 22 business agents for everything else. Arthur pressure-tests your idea. Kukai gets you found in search. Moneta models whether your pricing actually works.
I’m running everything. I’m running on empty.
You have 47 browser tabs open and none of them are the product. One is a terms of service template. Another is a pricing strategy blog post. Three are competitor websites. You’re not making progress — you’re maintaining the illusion of progress, because you’re one person doing seven jobs and finishing none of them.
Elven gives you 32 agents across 7 teams that actually coordinate. Galadriel orchestrates across teams. Arthur handles strategy. Turgon runs engineering. Hermes runs marketing. Censoria handles compliance. Moneta covers finance.
Same product. Same price. Different starting points. The team adapts to where you are.
HONEST COMPARISONS
You have options. Here's how they compare.
We're not going to pretend alternatives don't exist. Here's why people choose Elven anyway.
"I'll just build my own agents."
You absolutely can. But building 32 coordinated agents with shared memory, handoff protocols, operational modes, and 11 maturity models takes months. Elven is what that looks like when someone already did the work. It works alongside what you've already built — you're not replacing your setup, you're upgrading it.
"There are free repos with 600+ agents."
There are — and they're genuinely useful. For coding. But none will validate your business model, draft your terms of service, or plan your go-to-market. Elven isn't competing with those repos. They help you build the app. Elven helps you turn the app into a business. 32 agents. 22 of them aren't engineers.
"I just use ChatGPT / Claude directly."
That's one smart generalist. Elven is 32 specialists who coordinate. Your CEO agent doesn't write code. Your compliance agent doesn't do UX design. Specialization is the difference between "AI-assisted" and "AI-operated."
"I already use Cursor."
Great — keep using it. Cursor handles code. Elven handles the other 80%: strategy, marketing, content, finance, compliance, operations, and project management. They solve different problems.
"My company has an enterprise AI platform."
Those cost $500+/month, require a sales call, take weeks to configure, and are locked to your employer. Elven is $197 once, sets up in 10 minutes, and it's yours — for this job, the next job, and the side project.
"I'll figure out the business stuff later."
You won't. The graveyard of brilliant apps that never found customers is enormous, and the cause of death is almost never the code. Elven gives you the business scaffolding from day one. Not because you can't learn it eventually. Because "eventually" is where side projects go to die.
Start free. Think for $29. Ship for $197.
Starter
$29
Plan what to build. Challenge your assumptions. Create your own agents.
All paid plans include a 14-day refund guarantee. One-time purchase — no subscriptions.
REAL ACTIONS
Connected to your real tools. Not just generating text.
Your agents connect to real tools via Model Context Protocol. This isn't text generation. It's execution.
🏗️
Build your website
Manage pages, update copy, and ship changes through Framer MCP.
💻
Write and ship code
Access your codebase, create files, run tests, deploy.
🔍
Research the market
Search the web, analyze competitors, gather pricing data.
🌐
Run browser tasks
Automate workflows in Chrome via DevTools MCP.
🖼️
Create visual assets
Generate and edit SVGs, diagrams, and graphics with SVGMaker MCP.
📝
Manage your content
Update your CMS, draft blog posts, organize content.
⚡
Execute terminal commands
Build scripts, data processing, deployments.
🔗
Connect your tools
Webhooks, APIs, integrations wired together.
CAPABILITIES
What makes Claude Code agents a system, not a collection.
Individual agents are useful. What makes Elven different is what happens between them.
A runtime, not a template.
Four modes that change how your agents think. Execution for getting things done. Design for critical decisions. Research for objective analysis. Parallel for simultaneous workstreams. Switch with a single command. The same team, behaving differently depending on what the work needs.
Decisions that survive scrutiny.
One command switches your entire team into critical thinking mode. Every agent generates 3+ distinct approaches. Then Maat — the built-in devil’s advocate — stress-tests each one. You don’t ship the first idea that sounded good. You ship the one that survived structured opposition.
Intelligence without agenda.
Switch to Research Mode and your agents gather data without pushing recommendations. Competitive landscapes, market sizing, pricing benchmarks — collected and organized, with no conclusions attached. Good decisions start with clean data.
Your team gets smarter over time.
The context system gives all 32 agents shared knowledge that compounds. Market insights, technical constraints, customer patterns — captured once, available everywhere. Your AI team three months from now is sharper than your AI team today.
The right specialist, automatically.
Agents know their boundaries. When a marketing question comes up in an engineering conversation, the work gets routed to the right team. No manual switching. No copy-pasting between chats. The system handles coordination so you can focus on decisions.
Structure that scales.
Real sprint planning. Real checkpoints. Real retrospectives. Your agents don’t just respond to prompts — they execute within a methodology that tracks progress, manages scope, and keeps parallel workstreams from colliding.
YOUR TEAM
32 agents. 7 teams. One org chart.
Not a flat list of bots. A structured organization with reporting lines, handoff protocols, and shared memory. Each agent knows its role, its scope, and who to hand off to.
WHAT'S INSIDE
This is not a zip file of prompts.
It's a complete operating system for building with AI. Every piece is written, tested, and wired together.
32 Specialized Agents
Each with a full specification: role definition, responsibilities, collaboration rules, escalation paths, and example workflows.
42,000+ Words
Of documentation. Not filler — context engineering that makes each agent genuinely useful at its specialty.
4 Operational Modes
Execution, Design, Research, and Parallel — behavioral switches that change how your agents work, not just what they say.
62 Pre-Built Files
Ready to deploy. Agent definitions, protocols, templates, maturity models — the complete framework.
7 Coordinated Teams
Leadership, Engineering, Infrastructure, Growth, Content, Governance, and Management — structured like a real company.
11 Maturity Models
Every agent scored against quality frameworks. This is methodology, not a prompt pack.
PRICING
Try the talent. Get the strategy. Run the company.
Start free with 8 specialist agents. Add strategic planning for $29. Get the full 32-agent system when you're ready.
Free
forever
8 specialist agents. Production-ready. No strings.
✓ Cirdan — Backend architecture
✓ Legolas — Frontend and CLI
✓ Hermes — Growth marketing
✓ Brigid — Content strategy
✓ Kukai, Moneta, Arwen, Elladan
No signup required. Drop into your project and go.
Starter
$29
Stop guessing, start planning. Get a strategist to build your roadmap, a critic to stress-test it, and a builder to create agents for your domain.
✓ Strategic roadmap with milestones and projects
✓ Devil's advocate critique (Design Mode)
✓ Create custom agents tailored to your stack
✓ All 8 free agents included
✓ 3 operating modes: Execution, Design, Research
One-time purchase • instant delivery
Full Team
$197
When you're ready to go from planning to execution. All 32 agents, coordinated by Galadriel, running real sprints.
✓ Everything in Starter, plus 21 more agents
✓ Galadriel (COO) — Orchestration
✓ 8 protocols + 11 maturity models
✓ 4 operational modes + sprints
✓ Shared context + compound learning
14-day refund • $149 if upgrading from Starter • instant delivery
Recommended
| Feature | Free | Starter | Full Team |
|---|---|---|---|
| Specialist agents | 8 | 11 | 32 |
| Backend, frontend, marketing, content, SEO, finance, UX, prompts | ✓ | ✓ | ✓ |
| Strategic roadmap & Grand Plan | — | ✓ | ✓ |
| Devil's advocate critique (Design Mode) | — | ✓ | ✓ |
| Create custom agents | — | ✓ | ✓ |
| Operating modes | 1 | 3 | 4 |
| COO orchestration (Galadriel) | — | — | ✓ |
| Sprint methodology | — | — | ✓ |
| 8 protocols + 11 maturity models | — | — | ✓ |
| Shared context & compound learning | — | — | ✓ |
| 7 coordinated teams | — | — | ✓ |
QUESTIONS
Common questions, straight answers.
The stuff you're probably wondering about.
What do I get with Elven?
What's the difference between Free, Starter, and Full?
What if Claude Code adds native agent features?
Can I upgrade from Starter to Full?
What's your refund policy?
You've read this far because you're already building this.
The agents. The prompts. The workflows. You know where this is going. Start with 8 free agents, add the strategy layer for $29, or get the full system when you're ready.
Get the build notes
Agent techniques, real workflows, and what's actually working. From one builder to another.
No spam. Unsubscribe anytime.
WHAT'S INSIDE
This is not a zip file of prompts.
It's a complete operating system for building with AI. Every piece is written, tested, and wired together.
32
Specialized Agents
42K+
Words of Documentation
62
Pre-Built Files
7
Coordinated Teams
4
Operational Modes
11
Maturity Models