Project: AgentSaaSy

Watch the swarm work.
Then measure it.

AgentSaaSy is an R&D platform for enterprise AI Agent stacks. A test harness orchestrates AI Agents through real workloads while AEQ, the Agent Efficiency Quotient, scores the architecture itself: how much business value the design delivers per token it consumes. Specs do not predict adequacy. Only measuring the model-workload pair does.

AEQ=Business Value DeliveredTokens Consumed

The Swarm, Live

A simple workflow, end to end: a request enters, the harness routes it to a planner, fans work out to specialist AI Agents in parallel, a cross-family judge reviews the result, and the AEQ gate issues a pre-registered verdict before anything ships. Every hop is metered.

SIMULATED DATA · DEMO

speed

AEQ Monitor (simulated)

--BVD units / 1K tokens

GREEN

YELLOW

RED

Run Counters

requests completed

tokens metered

agents in swarm

green gate rate

Harness Log

Architecture Layers

Four layers do the work. One measurement plane cuts across all of them. Hover or tap a layer to inspect it.

AgentSaaSy Application Layer

Workflow definitions, business value rubrics, and the SaaS-substitution surface where AI Agent stacks replace seat-licensed software.

workflowsBVD rubricsoutputs

Harness & Orchestration Layer

The conductor. Routes requests, fans out parallel work, enforces pre-registered gate thresholds, and coordinates cross-family judging.

routerschedulergate logicrun manifests

Agent Layer

The swarm: planner, retriever, workers, and an independent judge from a different model family than the workers it reviews.

plannerretrieverworkersjudge

Model Layer

Frontier APIs and quantized local models, treated as interchangeable capacity. The pair (model + workload) is what gets measured, never the spec sheet.

Anthropic APIOpenAI APIquantized local

AEQ MEASUREMENT PLANE

Layer Detail

Select a layer

Hover over any layer on the left, or the amber AEQ plane, to see what it does and how it connects to the rest of the stack.

The Stack

What actually runs underneath the demo above.

Pythonharness + experiments

Anthropic APIagents + cross-family judge

OpenAI APIagents under test

tiktokenexact token accounting

Pre-registration docsgates declared before data

AEQ Grid3x3x3 certification cells

GitHub · ibucketbranchopen evidence trail

Canvas + vanilla JSthis visualization

From Simple to Swarm-Scale

This page shows one workflow. The architecture is built to grow.

NOW

One workflow, fully metered

Single request pipeline with parallel fan-out, cross-family judging, and a pre-registered AEQ gate on every run.

Multi-workflow swarms

Concurrent workflows sharing the agent pool, tiered model routing, and per-workflow AEQ baselines to catch architecture waste.

LATER

AEQ Grid certification

The full 3x3x3 grid: query classes by model tiers by repeated runs, producing GREEN / YELLOW / RED verdicts for model-workload pairs.

AgentSaaSy · an R&D project by Michael Valderrama · bucketbranch.ai

All meters on this page run on simulated demo data. Published claims use real API results only.

The measured version

Everything above this line runs on simulated data. The same architecture was put through a real workload and the result was written down: an enterprise asset management build answering at $0.0009 per query, with the run date and the pricing caveat on the record.

The enterprise asset management case study The technical white paper behind it The AEQ specification the score is defined by