— the action engine

The Glass Box Engine

An orchestrator that directs, expert agents that execute. This is what turns a request into real action — not just text.

— the orchestrator

It directs. It executes nothing.

Every request is read by an orchestrator whose only job is to choose among three routes: answer directly, send a scout for a read-only recon before deciding, or break the work into a structured plan handed to experts. It never touches write tools itself — it directs. That decision is semantic, made by the model in any language, never a hard-coded keyword list.

three routes, decided by meaning
request orchestrator direct answer | scout (read-only) | plan → experts
conversation → direct answer · simple check → a scout probes, then the orchestrator decides · complex task → a plan
— execution by waves

A team of experts, in parallel

The plan is executed by expert agents, each running its own reason → act → observe loop with real tools. Independent tasks run in parallel; each wave automatically inherits the previous one's results. No shared files needed.

wave 1 — parallel
expert · research
expert · data
expert · drafting
wave 2 — dependencies resolved
expert · compiles (gets the results)
wave 3 — if high effort
reviewer · coverage · consistency
→ final synthesis, delivered to the user
Orchestrator — depth 0 it steers and structures: a read-only scout for a quick check, or create_plan to delegate. No write action by itself.
The plan — split into parallel waves
Expert · wave 1reason → act → observe
Expert · wave 1reason → act → observe
Expert · wave 1reason → act → observe
↓ wave 2 inherits wave 1's outputs (context_from_parents)
Expert · wave 2compiles, dependencies resolved
Cascade — high effortan expert can open its own sub-plan
Reviewer — if high/max effortcoverage · consistency · sources → 1 targeted fix-plan if gaps
Final synthesisdelivered to the user
agentic flow — orchestrator → plan → ReAct experts in parallel waves
— the effort dial

You set how hard it thinks

From most economical to most rigorous. Higher effort raises the ceilings (depth, agent count, reasoning budget) and adds an automatic reviewer that re-reads every artifact — coverage, consistency, sources. The level guides; it never forces, never changes the guardrails, and stays bounded by an absolute cap (depth ≤ 10).

low

Direct answer

Fast, little decomposition.

depth 0 · 3 agents · reasoning off

medium · default

Balanced

Speed and quality.

depth 1 · 8 agents · reasoning low

high

Fine decomposition

+ quality-review pass.

depth 2 · 16 agents · reasoning medium · review

max

Critical iteration

Deep work, intermediate results challenged.

depth 3 · 30 agents · reasoning high · review

— agnostic core

The "judgment" decisions — should it delegate? is this fact durable? is this signal noise? — are delegated to the model through guidance, never hard-coded. The system works in any language, and the model provider is configurable per deployment.

— access

A use case in mind?

Tell us your use case — individual reply within 48h.