Pre-Seed · $300–500K SAFE · June 2026

DynaMind

Multi-agent AI-powered cryptocurrency trading platform. Intelligence infrastructure for financial AI.

GitHubX / TwitterPresale

The Problem

Financial AI is broken.

Everyone uses OpenAI APIs and calls it “AI trading.” No proprietary data layer. No specialized models. Single agents, single points of failure. 74–89% of AI trading bots fail within 90 days.

The market needs infrastructure, not another wrapper.

The Solution

DynaMind is building the intelligence infrastructure for financial AI. Six layers. Full vertical integration.

1

Terminal

The interface for financial workflows. Human oversight for autonomous agents.

2

Financial API

Structured market intelligence. Market data, signals, research objects, portfolio data. Developer-ready.

3

Embeddings

Proprietary financial embedding model. Search, memory, and RAG for financial agents.

4

Finance SLMs

Specialized reasoning models. Financial analysis, risk assessment, research, and agent workflows.

5

Autonomous Agents

28 specialized agents in a 3-tier hierarchy. Domain analysts, managers, orchestrator. Risk engine has final veto.

6

Hedge Fund

Our first customer. Proof that the stack works in live markets.

The Moat

Most startups have one thing. We have a technology stack.

Proprietary financial datasets (100+ sources)
Embedding model (finance-native)
Finance SLM research
RL environment (Fin-RL-Gym)
Price forecasting research (DiffuPrice)
Multi-agent architecture (28 agents, production)
Trading engine (risk-first, 100+ exchanges)
Terminal and API
Live execution
Years of iteration from Stacker AI (2023)

A competitor would need to build all of this to match where we are today. That is years, not months.

Key Metrics

28

AI Agents

100+

Exchanges

8

Crypto ETFs

~65%

Platform Complete

Three Revenue Paths

B2C

Trading Platform

28-agent system access for retail and prosumer traders. Subscription + performance fees.

B2B

Data API + DynaTrader

Financial builders and institutions build on our infrastructure.

Platform

License the Stack

We are building Path 3, proving it with Path 1.

The Ask

$300–500K pre-seed SAFE

3 months

  • Public terminal
  • API launch
  • First paying customers
  • Hedge fund running
  • Benchmark finance embeddings

6 months

  • Enterprise pilots
  • Finance SLM beta
  • API expansion
  • Growing developer ecosystem

12 months

  • Infrastructure revenue
  • Institutional pilots
  • Finance models
  • Series A readiness

Use of Funds

Engineering (solo to team)40%
Infrastructure30%
GTM & user acquisition20%
Legal & compliance10%

What's Live

  • Data layer: 100 sources ingesting, embeddings operational, AI news sentiment in production
  • Models: SLMs in training, RL gyms operational, custom algorithms in development
  • Infrastructure: 28 agents deployed, Telegram + dashboard live, risk engine active
  • 100+ exchange connections via CCXT
  • Agent feedback loop: results continuously train the system

Founder

Michiel Horstman

  • Built AI trading systems since 2023 (Stacker AI)
  • Built financial infrastructure before foundation models became agentic
  • Own embedding models, building finance SLMs
  • Production multi-agent architecture
  • Full vertical integration: data, embeddings, models, agents, execution, terminal
  • Solo technical founder capable of shipping the entire stack
michiel@4mlabs.comgithub.com/michielhdoteth/dynamind

Why Now

  • AI agents are mainstream. We are the financial execution layer.
  • 73% of autonomous trading failures are risk management gaps, not prediction errors. Our platform treats risk as a first-class concern.
  • LATAM crypto adoption is exploding (Mexico: 15M+ retail users, 60% YoY growth). Underserved by US VCs.
  • The window for financial AI infrastructure is now. First movers with proprietary data and models win.