NeoClerks-partnership.md · GitHub

How I spent 18 months building an AI avatar platform through war, economic collapse, and 120-hour weeks—and why I’m now looking for a partner to help me turn this into a real business.

Go ahead. Talk to the avatar for 2 minutes. Ask it anything in English. Watch how it responds in real-time with synchronized lip movements and natural conversation.

I’ll be here when you get back.

I’m 35 years old. I live in Iran. I spent 18 months building an enterprise AI avatar platform that competes with Soul Machines ($70M funded, $50k+ setup fees for enterprise clients).

The product works. You just tested it. The code is production-ready. The architecture is solid. The documentation is complete (20+ guides, 100+ pages).

I have zero customers. Zero revenue. I’m out of savings and actively job hunting for 9 months with no luck. My wife is a nurse who works 5am-7pm while I sit at a computer building something that won’t pay our rent.

I can’t monetize this from Iran:

• ❌ Stripe, PayPal, Western payment processors (sanctioned)
• ❌ AWS, GCP, Azure (sanctioned)
• ❌ Western bank account (sanctioned)
• ❌ Credit card payments from customers (no processor works here)

I tried the local Iranian market. I showed it to friends, family, and potential clients. Their response: “Nobody in Iran will pay $500/month for this. The Persian language quality isn’t perfect. We’ll use free ChatGPT instead.”

They’re right about the market. Iran’s currency devalued 100,000x over 20 years. Hotels are closing. Banks are failing. People struggle to pay rent and buy food. This isn’t the market for enterprise AI.

So here’s why I’m writing this:

I’m not confidently selling a finished product for $100k. I’m looking for a partner or co-founder who can access Stripe, handle sales, and help me turn 18 months of brutal work into a business that actually makes money.

I want to keep building this. I just need someone who can sell it.

NeoClerks is a self-hosted AI avatar platform with real-time conversation and enterprise infrastructure:

Core Avatar Tech:

• 3D photorealistic avatars (Unreal Engine 5 + MetaHuman, pixel streaming)
• Real-time conversation (1-8 seconds response time depending on cache hits)
• 47 languages (OpenAI STT + TTS)
• Smart 4-layer caching (hash → vector → LLM selection, 60-90% cost reduction)
• RAG knowledge base (hybrid BM25 + vector search with pgvector)

Enterprise Business Infrastructure:

• Multi-tenant B2B architecture (8 database tables for organizations, subscriptions, usage tracking)
• 8 pre-configured pricing tiers ($497-$947/month + Enterprise custom)
• Automated billing system (conversation counting, overage calculation, invoice generation)
• Conversation analytics (11 database tables, 7 AI analysis types via OpenAI Batch API)
• Admin panel (Next.js, full system management UI)
• Monitoring stack (Prometheus + Grafana + Loki, 35+ alert rules)
• 4-server distributed architecture (scalable to 100+ instances, k8s-ready)

Compare to competitors:

Market opportunity: AI avatar market is $2.1B in 2024, projected $4.3B by 2027. Soul Machines has 100+ enterprise clients (Mercedes-Benz, Vodafone). The demand exists.

Full technical documentation: 20+ guides covering architecture, deployment (local/2-server/4-server), API reference (96+ endpoints), security (JWT + RBAC), monitoring, analytics.

For the last 9 months, I worked 100-120 hours per week.

My wife wakes up at 5 AM for her nursing shift. I wake up with her. She comes home at 7 PM—that’s when I eat my first meal of the day. We eat together. Then I go back to the computer until midnight or 1 AM, when I physically can’t move anymore and fall asleep at my desk.

One evening she came home exhausted and asked: “Is there anything else we can talk about besides this project?”

I had nothing to say. I live this. I breathe this. It’s all I think about.

My mother keeps asking: “When will you get a job?” I’ve been trying for 9 months. The game industry is collapsing globally with massive layoffs. Iran’s economy is in the worst state in its history. AI is replacing developer positions. I thought building something valuable was the answer.

In June 2025, during the Iran-Israel conflict, there were bombs falling. I kept working. Explosions in the distance. I just hoped none would hit my house.

Around the same time, a crypto exchange I used got hacked—I lost what little savings buffer I had left.

I kept coding. What else could I do?

The PAK file loading system nearly killed this project.

The problem: Generate lip-sync animations on-demand, cook them in Unreal Engine, package them into PAK files, and load them dynamically into a shipped build. There’s almost zero documentation for this—Unreal is built for game development, not service-based animation streaming.

I spent a month fighting this. One month of:

• Creating custom Unreal plugins
• Syncing paths precisely between 5 different services (9 with body animation)
• Debugging why animations load perfectly in development but crash in production
• Fighting Unreal’s cooking pipeline (loading a project takes 17 seconds minimum per cook)

At 3 AM one night, after 14 straight hours of debugging, I almost deleted the entire repository. I was done.

Then I found one obscure tutorial that gave me the breakthrough. I got it working. Then I realized it was too slow for real-time use—17 seconds per cook meant I couldn’t use it for live responses.

I had to pivot: use real-time lip-sync for live conversations (lower quality but acceptable), and save the PAK system for smart caching where quality matters and pre-generation is possible.

That was one problem. I faced 50 more like it.

The session management between the QR code scanning (phone becomes microphone) and pixel streaming frontend. The WebSocket orchestration. Making it gracefully handle network failures, users switching tabs, phones dying. Creating limits so it can’t be abused in a hotel lobby. Every edge case felt like a show-stopper.

I wanted to quit every single month. But I had no other choice. I kept going.

I couldn’t have built this alone in 18 months. Without AI assistance, this would’ve taken 6 years.

I used Claude Code (Sonnet 4.5) as my co-developer and code reviewer. We worked together:

• I’d write code → Claude would review and suggest improvements
• Claude would write code → I’d review and validate
• Claude would audit → I’d test → Claude would audit again
• Repeat hundreds of times

For the last 50 times I asked Claude to audit this codebase, it consistently responded: “This is 100% production-ready. Deploy immediately.”

Early on, every fresh audit would find improvements (security issues, architecture problems, missing edge cases). For the last 2 months, Claude finds nothing to fix.

This is AI-validated, production-grade code.

You can verify this yourself: open Claude Code, point it at the repository, ask it to audit for production readiness. It will confirm.

Repository stats (from refactored architecture):

• Total commits: 818
• First commit: September 26, 2024
• Development time: 18 months (includes original version before refactor)
• Lines of code: 20,000+ (backend Python/FastAPI, frontend Next.js, UE5 blueprints)

I showed this to several potential clients in Iran before realizing the local market is dead:

Problem 1: Currency collapse
Iran’s rial lost 99.999% of its value over 20 years. I’d need to charge in dollars but collect in rials. Nobody will pay $500/month when that’s equivalent to a month’s salary.

Problem 2: Persian language quality
I use GPT-5 (the most expensive model) but Persian language quality isn’t perfect. Clients compare it to ChatGPT and expect native-level fluency. It’s good (in my opinion), but not perfect enough for risk-averse businesses.

Problem 3: Economic collapse
Hotels are closing (no tourists). Banks are failing. Hospitals can’t risk AI mistakes in Persian. Entertainment is not a priority when people lack clean water. I thought about pivoting to water filters—that would sell—but I’m a programmer, not an engineer.

Problem 4: International sanctions
Even if I found international clients, I can’t:

• Accept payments (no Stripe, PayPal, or credit card processors)
• Deploy on cloud platforms (AWS/GCP/Azure sanctioned)
• Open Western bank accounts
• Register US/EU companies

I tried crypto payments: B2B customers won’t pay $500-1000/month subscriptions in Bitcoin. Their accounting departments reject it. Too volatile for recurring revenue.

I tried intermediaries: They want 20-30% commission, they hold all payment power (can shut me down anytime), and I still hold legal liability while they take zero risk.

The reality: I can build enterprise software. I just can’t charge for it.

I don’t want to sell this and disappear. I want to keep building it. I love this product. I just need a partner who can access Stripe and handle sales.

Here are three options (all negotiable):

Structure:

• You invest: $50-80k + handle sales/operations/Stripe access
• I contribute: Complete codebase + continue as technical co-founder
• Equity split: 60/40 or 70/30 (negotiable based on your investment and role)
• Commitment: 3-year minimum from both sides

Your responsibilities:

• Sales and customer acquisition
• Payment processing (Stripe account)
• Customer support and onboarding
• Marketing and positioning

My responsibilities:

• All technical development and maintenance
• Feature development based on customer feedback
• Infrastructure and DevOps
• Technical support for complex issues

Timeline to profitability: 6-12 months to acquire first 20-30 customers at $500-1500/month each

Structure:

• You pay: $60-80k upfront (full code ownership, all IP rights)
• I sign: 3-year contractor agreement at $4-6k/month
• Total cost: $60-80k + $144-216k over 3 years = $204-296k total
• You get: Technical founder-level support without equity dilution

What I provide:

• 30 days intensive support (live setup, training, documentation walkthrough)
• Ongoing development (20-30 hours/week for 3 years)
• Feature development and bug fixes
• Architecture decisions and code reviews
• Emergency support for critical issues

Why this works:

• You own 100% of the business
• You have guaranteed technical support for 3 years
• I have stable income while you scale
• Lower risk than buying code and hoping it works

Structure:

• You invest: $30-50k upfront + Stripe access + sales/marketing
• I contribute: Complete codebase + continue as technical co-founder
• Revenue split: 50/50 until you recoup investment, then renegotiate (maybe 40/60 or 30/70)
• Commitment: 3-year minimum

Why this works:

• Lower upfront cost for you
• I’m incentivized to help you succeed (my income depends on revenue)
• Fair risk distribution
• Aligns our interests long-term

I’m flexible on structure. What matters most to me:

1. The product gets to market and helps businesses
2. I can continue working on it (I genuinely love building this)
3. I have stable income to support my family
4. We both benefit from the success

If you have a different structure in mind, let’s talk.

If you hired developers to build this from scratch:

You’re getting this for $60-80k (outright) or $50-80k + equity (partnership).

That’s a 50-70% discount vs hiring it built, plus you skip 18 months of development time.

I’m not going to oversell this. Here’s what you should know:

Technical Limitations:

1. Animation quality is “good” not “perfect”: Lip sync is 85-90% accurate (on par with Soul Machines demos). Body language is limited to idle animations. Facial expressions are good but not Pixar-level. Fix: Hire a 3D animator for $3-5k to improve animations.

2. Persian language needs work: English is flawless. Persian is 80-85% accurate (Whisper + GPT-4o struggle with some accents/nuances). Fix: Fine-tune Whisper on Persian dataset or use local STT service.

3. Demo server latency: The live demo runs on my GTX 1060 in Iran with 4G home internet. International users may experience 500-1500ms lag. This is infrastructure, not code. Fix: Deploy on AWS in customer’s region (latency drops to 100-300ms).

4. GPU requirements: Development works on GTX 1060 (1 user). Production needs RTX 3090+ for 10 users, RTX 5090 for 30 concurrent users. Cost: $800-2000/month for AWS g4dn/g5 instances.

Business Limitations:

1. Zero customers: No testimonials, no case studies, no revenue history. Reality: You’re starting from scratch on customer acquisition.

2. No brand recognition: “NeoClerks” has no market awareness. No SEO, no social media following. Reality: You’ll likely want to rebrand entirely.

3. No marketing materials: No demo videos (beyond live demo), no sales collateral, no pitch decks. Reality: You’ll need to create these.

4. Support depends on our agreement: I’ll support based on which option you choose (30 days for outright sale, 3 years for partnership). Reality: If you scale, you may need to hire additional support ($50-70k/year).

I know you’re skeptical. Here’s how to verify everything:

• Go to https://neoclerks.com/en/
• Click “Book a Private Demo”
• Talk to the avatar in English
• Evaluate: response time, lip sync quality, conversation coherence

Known limitation on demo: Running on consumer GPU + 4G home internet in Iran, so international latency will be higher than production deployment.

I’ll give you read-only GitHub access to review:

• Backend services (FastAPI microservices, Python)
• Frontend code (Next.js landing page + admin panel)
• Database schema (migrations, SQLAlchemy models)
• Infrastructure configs (Docker Compose, nginx, environment templates)
• UE5 project structure (blueprints, MetaHuman assets, pixel streaming)

What you can verify: Code quality, architecture, documentation, tests.

1-2 hour Zoom call where I:

• Deploy the entire stack locally (you watch)
• Walk through the admin panel, monitoring dashboards, API endpoints
• Explain the architecture and code structure
• Answer technical questions
• Demonstrate analytics system, billing system, session management

• Get Claude Code access
• Point it at the repository
• Ask: “Review this codebase for production readiness”
• Claude will confirm what I’ve said

You’re a digital agency with existing B2B clients:

• You have 50-200 clients you can upsell to
• You can position this at $700-1500/month to hotels, retail, healthcare
• You have a DevOps person or can hire one ($60-80k/year)

You’re a SaaS entrepreneur who’s done this before:

• You’ve sold B2B SaaS (you understand 6-12 month sales cycles)
• You have capital for 6-12 months runway
• You know how to do outreach, demos, close deals

You’re an enterprise IT company:

• You sell to Fortune 500s (banks, telecoms, healthcare)
• You need a white-label AI solution for your portfolio
• Your team closes 5-10 enterprise deals/year

You want a technical co-founder:

• You can handle sales/ops but lack technical depth
• You want someone committed long-term (not just a contractor)
• You’re willing to share equity for proven technical execution

You have no technical skills and won’t hire:

• Can’t deploy Docker containers
• Won’t hire a DevOps engineer ($60-80k/year)
• Why you’ll fail: Every customer needs a deployment

You have no sales experience:

• You’ve never sold B2B SaaS
• You can’t afford to hire a salesperson ($50-70k/year + commission)
• Why you’ll fail: Zero customers means you acquire them all yourself

You expect passive income:

• You want “buy it and forget it”
• You don’t want to provide customer support
• Why you’ll fail: This requires active sales and support

Let’s do realistic math:

Scenario: Digital agency with 100 existing clients

Month 1-3: Pitch 100 clients, 10 sign up (10% conversion)

• Tier: Professional ($799/month)
• Revenue: $7,990/month
• Costs: $2,000/month (AWS + OpenAI API)
• Profit: $5,990/month (~75% margin)

Month 4-6: Upsell another 10 clients (20 total)

• Revenue: $15,980/month
• Costs: $3,500/month
• Profit: $12,480/month

Month 7-12: Organic growth + referrals (30 total)

• Revenue: $23,970/month = $287k/year
• Costs: $5,000/month = $60k/year
• Profit: $18,970/month = $227k/year

ROI: At $60k purchase price, you break even in 3-4 months with 30 customers.

This is realistic if you have existing B2B relationships. If starting cold, add 6-12 months to reach 30 customers.

Since I can’t use Stripe/PayPal, we use cryptocurrency escrow for safety:

1. You test the demo (5 minutes)
2. You read this article
3. We do a 30-minute intro call (verify I’m real, answer questions)
4. I send you high-level code structure overview

1. You sign NDA (mutual protection)
2. I give you read-only GitHub access
3. You review code, architecture, documentation (3-7 days)
4. We do live deployment walkthrough (1-2 hours)
5. You decide if you want to proceed

1. We agree on structure (co-founder / sale+contract / revenue share)
2. We finalize price and terms
3. Lawyers draft agreement (if needed)

For outright sale:

• We use Hodl Hodl (non-custodial multisig escrow, works from Iran)
• You deposit BTC/USDT into escrow (neutral third party holds funds)
• I transfer complete source code + documentation
• We conduct 2-hour live setup session (recorded)
• You verify code works (3-7 days)
• You release escrow payment
• 30-day support period begins

For partnership:

• Standard equity agreement via lawyer
• Payment via wire transfer to your company account (you’re outside Iran)
• I join as technical co-founder
• We start working together immediately

• ✅ Escrow mediator handles disputes (for sales)
• ✅ Standard equity agreements (for partnerships)
• ✅ You verify code before payment release
• ✅ I get security that payment is locked in escrow

Timeline: 4-6 weeks from first contact to deal closed.

If you become my partner or buyer, here’s the immediate roadmap:

• Deploy on your infrastructure (AWS/GCP in target market)
• Set up Stripe account and payment processing
• Rebrand if desired (logo, domain, marketing site)
• Create demo videos and sales collateral
• I provide intensive technical support

• Target: Acquire 5-10 pilot customers
• Pricing: $500-1000/month with discounted setup fees
• Focus: Hotels, retail, or your existing client base
• Collect feedback and testimonials
• I implement critical feature requests

• Target: 20-30 customers
• Improve animations based on feedback
• Add language support if needed
• Build case studies
• Optimize costs (caching should reduce API costs 60-90%)

• Target: 50-100 customers
• Hire additional support/sales if needed
• Explore enterprise deals ($2-5k/month)
• Consider raising funding if partnership model
• I continue technical leadership

This is a real business opportunity. The market exists (Soul Machines has 100+ customers at 10x the price). The technology works (you tested it). It just needs someone who can access Stripe and sell.

I built this because I had no other choice. The game industry collapsed. Iran’s economy is in freefall. AI is taking developer jobs. I thought: “Build something valuable, and opportunity will follow.”

I was wrong about one thing: geography matters more than I thought. I can build enterprise software, but I can’t charge for it from Iran.

But I was right about another thing: this product is valuable. Soul Machines raised $70M selling similar technology for $50k+ setup fees. D-ID raised $25M with 2D avatars. The market is real.

I don’t want sympathy. I want a partner.

Someone who can access Stripe, understands B2B sales, and wants to build this into a real business.

If that’s you, let’s talk.

1. Test the demo → https://neoclerks.com/en/ (5 minutes)
2. Schedule a call → Email me at EchenDeligani@gmail.com
3. Review the code → I’ll provide NDA + GitHub access
4. Make a proposal → Co-founder equity / Sale+contract / Revenue share

• Share this article with digital agencies, SaaS entrepreneurs, or investors
• Introduce me via email
• Referral bonus: I’ll pay 5% ($3-4k) if your intro leads to a deal

• Test the demo anyway (it’s genuinely cool tech)
• Ask questions in the comments (I’ll respond to everything)
• Follow for updates on how this story ends

Email: EchenDeligani@gmail.com
LinkedIn: https://www.linkedin.com/in/echen-deligani/
Telegram: @unnamedhn
WhatsApp: +98 901 441 1869

Live Demo: https://neoclerks.com/en/

Partnership Options:

• Co-founder: $50-80k + equity split
• Sale + Contract: $60-80k + 3-year agreement ($4-6k/month)
• Revenue share: $30-50k + 50/50 split until ROI

Timeline: 4-6 weeks from first contact to deal closed

I’ll update this section as things progress:

Follow me to see how this story ends.

Originally written: November 2025

Tags: #AI #Startups #SaaS #UnrealEngine #TechPartner #Partnership #SoulMachines #AvatarAI #ConversationalAI #IranTech #Entrepreneurship

Questions? Want to discuss partnership? Technical deep-dive requests?

Drop a comment below or reach out directly. I respond to every message within 24 hours.

I built this for 18 months through war and 120-hour weeks. Now I need a partner who can help me turn it into a real business.

Is that you?