SET
Give it a spec. Get merged features.
Autonomous multi-agent orchestration for Claude Code. We don't prompt — we specify. Every change is planned through OpenSpec, verified by quality gates, and merged automatically.
set-project init --project-type web From spec to merged code — fully autonomous
Multiple Claude agents in isolated git worktrees. Real branches, real merges.
Test, build, E2E, lint, review, spec coverage, smoke. Exit codes, not vibes.
3-layer templates + set-compare scoring. 87% structural overlap on micro-web, 75% on minishop. Convention compliance: 100%.
Structured artifacts: proposal → design → spec → tasks → code. No hallucination.
Issue pipeline: detect → investigate → fix → verify. Sentinel diagnoses before acting.
Core stays abstract. Web, voice, fintech — pluggable project types.
Review findings become rules. Gate failures teach the next run. Gets better with use.
Figma MCP → design tokens → Tailwind classes injected into agent context.
Real-time monitoring. Phases, tokens, agents, issues, learnings. Start from browser.
Hook-driven cross-session recall. Agents learn from each other. Shared across worktrees.
3-tier decision model. Crash recovery, checkpoint handling, stall investigation. 30s detection.
Multi-agent messaging. Broadcast status, avoid file conflicts, coordinate dependencies.
Waterfall took 8 months. This takes 8 hours.
The principle hasn't changed: output quality depends on input quality. A detailed spec used to mean months of upfront planning. Now it means hours of orchestrated agents building exactly what you described.
You are the product owner. The agents are the dev team. The spec is the sprint backlog. The better the spec, the better the result.
Business requirements, acceptance criteria (WHEN/THEN), technical constraints, dependency listing, seed data conventions.
Framework boilerplate, build config, test setup, linting rules, conventions. You say what. Templates handle how.
We treat determinism as an engineering problem
Run the same spec twice — we measure the structural overlap. 14 runs, 3 project types, set-compare scores every pair.
| challenge | approach | result |
|---|---|---|
| output divergence | 3-layer template system + set-compare | 87% micro-web · 75% minishop · 3 project types |
| convention compliance | route groups, colocation, naming rules | 100% across all runs |
| quality roulette | 7 programmatic gates (exit codes) | deterministic |
| hallucination | OpenSpec artifacts + acceptance criteria | spec-verified |
| spec drift | coverage tracking + auto-replan | 100% coverage |
| failure recovery | issue pipeline (detect → diagnose → fix) | auto-recovery |
| agent amnesia | hook-driven memory (infrastructure) | 100% capture |
Build your own project type
SET ships with a public web project type. The real power is building your own.
modules/web/
Next.js, Playwright, Prisma. 14+ orchestration runs across 3 projects. Divergence measured, conventions enforced.
your_project_type/
IDOR checks for fintech. HIPAA for healthcare. Your gates, your conventions. pip-installable plugin.
A real agent session — spec to merged code
Single-agent was the start. Orchestration is the present. Enterprise is preparing.
Systems like SET can do the work of a full development team — given the right spec and properly developed project types. This is the present, not the future.
Don't blame the model. 90% of agent failures are underspecification on our side. SET exists to enforce structure, verify output, and close those gaps.
Enterprise is next. On-premise models, secure multi-tenant — the infrastructure is coming. Every organization should prepare now.
Model providers will build orchestration natively. We welcome that. But we're not waiting.
when orchestration gets intense, defend your changes.
arrow keys + space. every change is a ship.