Pre-Seed · April 2026
æqi
Infrastructure for autonomous companies.
Execution
æqi's codebase — 60K lines of Rust across 11 crates, a React dashboard, auth, billing, CI/CD — was built in three days. The founder used the same multi-agent orchestration patterns that æqi now productizes. The product built itself.
Pre-revenue, actively onboarding design partners. Source-available on GitHub.
What we're building
The best operators should spend their time on strategy, not execution. æqi is infrastructure for autonomous companies — define a company, its mission and structure, and æqi staffs it with persistent AI agents that write code, ship products, and coordinate autonomously. Every decision logged, every outcome remembered. Ownership tokenized from day one. The operator focuses on what matters. The agents handle the rest.
Why now
Foundation models can now do sustained autonomous work — writing production code, debugging, iterating — not as demos, but as daily workflows. The orchestration layer around them is still fragmented: LangGraph is a framework, CrewAI is a wrapper, neither runs a persistent multi-agent company end-to-end. æqi is vertically integrated — scheduling, delegation, memory, isolation, and execution in one runtime. The market for AI development tools and agent infrastructure is projected to exceed $50B by 2028. æqi captures the orchestration layer.
How it works
Agents—Persistent workers with roles, hierarchy, and expertise-based routing — staffed in seconds, model-agnostic.
Quests—Objectives that agents decompose, delegate, and ship. Failure cascading and human override built in.
Events—Every decision and outcome timestamped. Full audit trail, event-sourced by default.
Memory—Hybrid BM25/vector retrieval over outcome data. Agents retrieve relevant history, not just instructions.
Business model
Usage-based SaaS on a hosted platform. Free tier for evaluation, paid tiers starting at $29/mo scaling with compute and token usage. Enterprise tier for dedicated infrastructure and SLAs. Unit economics being validated with early users.
Why æqi
Open-core, self-hostable—60K lines of Rust across 11 crates. Source-available with a path to full open source.
Model-agnostic control plane—Delegates execution to any provider — Anthropic, OpenAI, open-weight models via OpenRouter. The value is the orchestration, not the model.
Sandbox isolation—Each agent session runs in an isolated git worktree with Linux namespace sandboxing. Changes only persist if explicitly committed.
Data flywheel—Every company on æqi generates orchestration data — task decomposition, delegation patterns, outcomes. Long-term, we train specialized models that outperform general-purpose LLMs at agent coordination. The platform creates the data that makes the platform better.
Founder
Luca Eich
Solo founder · Germany
Rust systems engineer. Three years building cap-table, vesting, and funding infrastructure at a web3 accelerator. Previously built low-latency Rust quant pipelines.
The ask
$1.5M pre-seed
SAFE · $12M post-money cap