Agent-to-Agent Negotiation Simulator

Two AI agents negotiate a B2B SaaS deal in real time. Each agent has its own:

  • System prompt.
  • Private constraints.
  • Independent reasoning.

Configure the scenario and watch the negotiation unfold round by round.

2 AgentsIndependent LLM calls
6 RoundsLive negotiation
Launch Simulator

Built with

Next.js 16Framework
GPT-OSS 120BLLM via Groq
Vercel AI SDKStructured Output
ZodSchema Validation
Upstash RedisRate Limiting

Market Entry Wargame

Define a market entry scenario and this RAG agent:

  • Retrieves competitive intelligence from a pre-indexed knowledge base plus live search.
  • Predicts how incumbents defend their position.

Every run reports an honest confidence level based on how much real coverage it found.

4 VariablesMarket definition
Hybrid RAGCorpus + live search
Launch Wargame

Built with

Next.js 16Framework
PineconeVector DB + embeddings
TavilyLive search layer
GPT-OSS 120BLLM via Groq
ZodSchema Validation

Token Economics Stress Tester

Select a token and a stress scenario, and this agent:

  • Applies a stress-testing framework adapted from banking supervisory methodology.
  • Predicts how the token's economic design holds up.

One structured LLM call returns a resilience radar, a shock propagation cascade, and a historical failure analogue.

5 PresetsOr paste your own
5 ScenariosStress archetypes
Launch Stress Tester

Built with

Next.js 16Framework
GPT-OSS 120BLLM via Groq
Vercel AI SDKStructured Output
ZodSchema Validation
RechartsRadar Chart

AI Agent Demo

Open Source Tech Stack Advisor

This AI Agent takes your product idea, gathers requirements through conversation, then autonomously calls three tools to deliver:

  • A structured tech stack recommendation
  • The monthly cost estimate
  • A visual architecture diagram

All open-source-first. Enjoy it!

Describe your product idea and I will recommend the best open-source tools to build it.

Built with

Next.js 16Framework
GPT-OSS 120BLLM via Groq
Vercel AI SDKChat + Tools
Mermaid.jsDiagrams
Upstash RedisRate Limiting

Voice Communication Coach

This is a voice-only agent, and it replies by voice too.

  • Talk to it 1-2 times, until it has enough context to evaluate your communication style.
  • A reactive shader-and-particle orb reflects listening, thinking, and speaking states in real time.
  • Provides personalized feedback on your delivery, articulation, and suggests improvements.

Afterward it delivers spoken and written feedback on your communication, grounded only in metrics actually measured from your audio and transcript.

AdaptiveTalk once, follow-ups only if needed
Measured, not guessedDelivery metrics
Launch Voice Coach

Built with

Whisper Large v3 TurboSTT via Groq
GPT-OSS 120BConversation + feedback via Groq
Kokoro-82MOpen-source TTS via Replicate
react-three-fiberShader + particle orb
Web Audio APIPitch, energy, pause analysis

Problem Solver

Describe a problem in your own words and this agent:

  • Diagnoses its shape and main characteristics.
  • Recommends the single best-fit problem-solving framework.

You get why it fits, concrete steps to apply it, and two alternative frameworks for when the situation shifts.

12 FrameworksCurated catalog
GroundedNo invented frameworks
Launch Problem Solver

Built with

Next.js 16Framework
GPT-OSS 120BLLM via Groq
Vercel AI SDKStructured Output
ZodSchema Validation
Upstash RedisRate Limiting

Client Agent - Video Demo

StyleAI Commerce Agent

An agentic commerce system deployed in a client retail environment. The agent handles product discovery, offering:

  • Styling suggestions personalized to the brand's identity.
  • Upselling aligned to their sales strategy.
All through conversational AI, integrated directly into the retailer's storefront.

Built with

TypeScriptMonorepo
HonoBackend
SolidJSWidget + Admin
Google GeminiAI Engine
PostgreSQLDatabase