Open Source · Apache 2.0 · Protocol-Based

Unleash the Swarm.
Keep the Leash.

Conductor decomposes goals into capability-ordered plans, dispatches a swarm of specialized agents, and keeps humans in control at every critical junction. Infinite capabilities. 5 autonomy levels. Zero trust required.

conductor orchestration
Orchestrating
Goal

"Research Q4 financial regulations and prepare a compliance summary with recommendations"

Active Swarm
Researcher

Matching capabilities...

Planner
Coder
Reviewer
Deployer
Capabilities
5
Autonomy Levels
459
Tests Passing

Most AI Orchestration Is Fire and Forget

You tell a model to "do the thing." It tries. It fails. You debug the prompt. Nobody knows what happened, who approved it, or whether it will do something different tomorrow.

Without Conductor
One model handles everything
No idea why it chose that approach
"Just trust the AI" for critical decisions
Static prompts that never improve
Chat logs as your audit trail
With Conductor
42 specialized capabilities matched to every task
Topological plans with dependency ordering
5 autonomy levels: risk-proportionate human oversight
Adaptation engine that learns, with safety gates
Every decision Capsule-sealed and hash-chained
Autonomy Calculator

Five Levels. You Choose the Leash Length.

Every task gets an autonomy level based on its risk, the tools it uses, and your YAML-configurable rules. High-risk operations always stop for human approval.

Full Human ControlFull Autonomy
L3

CHECKPOINT_END

Final output reviewed.

Default. Agent executes autonomously, human reviews the final output before delivery.

42 Capabilities. 9 Domains.

Each capability is a YAML file that defines what an agent can do, what tools it needs, its risk level, and its dependencies. Drop a YAML file in the registry to add a new one. No code changes.

10

Analysis

Financial, market, risk, pattern detection, legal, accounting

4

Audio

Transcription, understanding, synthesis, diarization

2

Communication

Stakeholder updates, progress reporting

5

Creation

Code generation, tests, content, documentation, formal documents

4

Creative

Ideation, brainstorming, narrative design, concept generation

5

Cybersecurity

Threat analysis, vulnerability assessment, penetration testing, incident response

5

Data

ETL, transformation, quality, pipeline, warehousing

4

Education

Curriculum design, assessment, tutoring, learning analytics

5

Engineering

Systems design, optimization, simulation, prototyping

4

Environmental

Impact assessment, sustainability, resource monitoring, climate analysis

2

Execution

Deployment, shell execution

5

Governance

Policy enforcement, compliance monitoring, audit, risk management

5

Investigation

Research, debugging, root cause, vision, performance analysis

4

Linguistics

Translation, sentiment analysis, summarization, entity extraction

5

Mathematics

Statistical analysis, optimization, modeling, proof verification

5

Medicine

Clinical analysis, literature review, protocol design, diagnostic support

4

Operations

Workflow optimization, scheduling, resource allocation, monitoring

4

Philosophy

Ethical analysis, epistemology, logic, critical reasoning

4

Planning

Strategic, project, resource, task decomposition

5

Science

Hypothesis generation, experiment design, data analysis, literature synthesis

5

Social Science

Survey design, behavioral analysis, demographic modeling, policy analysis

5

Strategy

Competitive analysis, market entry, growth planning, scenario modeling

5

Sysadmin

Health check, diagnosis, remediation, validation, learning

5

Validation

Code review, compliance check, evidence collection, testing, verification

capabilities/registry/financial_analysis.yaml
name: financial_analysis
domain: analysis
description: Analyze financial data and regulations
risk_level: medium
required_tools:
- web_search
- document_reader
dependencies: []
estimated_duration_minutes: 15
7 Subsystems

From Natural Language to Governed Execution

01

Goal Analyzer

Natural language to capability requirements with intent classification

02

Capability Composer

Requirements to dependency-ordered execution plan via topological sort

03

Autonomy Calculator

Risk and confidence analysis to determine the right level of human oversight

04

Orchestrator

Multi-agent dispatch with parallel execution and Capsule-sealed decisions

05

Checkpoint Manager

Human approval gates with async wait patterns for high-risk operations

06

Adaptation Engine

Signal capture, pattern mining, and safety-validated self-improvement

07

Drift Detector

Behavioral consistency monitoring via cosine distance on embedding fingerprints

Safety Gates

Power Without the Danger.
Hardcoded, Not Configurable.

Some rules should never have an off switch. These constraints are compiled into the safety gate layer, not loaded from config files. You cannot accidentally disable them.

HIGH-risk tasks never auto-approved

Deploy, delete, and decision tasks always require human review. Hardcoded, not configurable.

Security agents never adapted

Auditor and deployer agents cannot receive prompt modifications from the adaptation engine.

Injection patterns always blocked

Shell injection, SQL injection, path traversal, and secret leakage blocked at the safety gate.

Kill switch on every iteration

Every agent checks the kill switch before each iteration. Once killed, it stays killed.

# Adaptation proposal for deployer agent
proposal = AdaptationProposal(
agent_type="deployer",
modification="prompt_update",
)
Blocked by Safety Gate

Security agents cannot receive prompt modifications

Not Another Prompt Chain

Conductor is a governed orchestration engine, not a wrapper around sequential LLM calls.

Task decomposition
Them:Manual prompt chaining
Conductor:Automatic capability-ordered plans
Minutes to plan, not hours
Agent selection
Them:One model for everything
Conductor:Infinite capabilities via synthesis
Right specialist, every time
Human oversight
Them:All or nothing
Conductor:5 autonomy levels (L0-L4)
Risk-proportionate control
Audit trail
Them:Chat logs
Conductor:Capsule-sealed decisions
Cryptographic proof
Self-improvement
Them:Static prompts
Conductor:Adaptation engine with safety gates
Gets better, stays safe
Drift detection
Them:Hope for the best
Conductor:Cosine distance monitoring
Catch problems before users do
pip install qp-conductor

Goal to Governed Plan in 10 Lines

quickstart.py
from qp_conductor import (
    GoalService, TaskService, InMemoryStorage,
    CapabilityRegistry, CapabilityComposer,
)

storage = InMemoryStorage()
goal_service = GoalService(storage=storage)

# One line: natural language to structured goal
goal = await goal_service.create(
    user_id="you",
    description="Research Q4 regulations and write a summary",
)

# Compose a capability-ordered execution plan
registry = CapabilityRegistry()  # 42 YAML capabilities
composer = CapabilityComposer(capability_registry=registry)

# Approve tasks from the priority inbox
task_service = TaskService(storage=storage)
tasks, count = await task_service.get_pending_review_tasks("you")
await task_service.batch_approve([t.id for t in tasks], "you")

Part of the Quantum Pipes Stack

Each independently useful. Together, the governed AI platform for autonomous organizations.

Unleash the Swarm.
Keep the Leash.

Infinite capabilities. 5 autonomy levels. Capsule-sealed decisions. Governed orchestration for autonomous organizations.

Apache 2.0 Protocol-Based Architecture Air-Gap Ready 459 Tests