The Rationale Engine for AI.
Govern, understand, and explain every decision. Visualize logic and align agent actions. An empirical safety infrastructure for autonomous systems.
➜ ~ git clone https://github.com/shansimmons-eng/tropelex.git
➜ ~ cd tropelex
➜ ~ pip install -r requirements.txt
Memory & Context
Accumulate knowledge across projects so sessions never start from scratch. Store decisions, patterns, and preferences in a unified graph.
Safety & Alignment
Immutable audit trails and silent drift detection. Ensure code never contradicts recorded decisions without explicit authorization.
Research & Ingestion
Deep multi-source insights. Scan web, code, and socials in parallel, synthesizing findings directly into citation-backed knowledge bases.
Everywhere You Build
Seamlessly integrated across web dashboards, IDE extensions, CLI tools, and agent protocol servers.
20 Core Engine Capabilities
Click any feature card to expand full technical specs, architecture mechanisms, and operational bounds.
Persistent Memory Graph
Retains decision patterns and context across projects so autonomous sessions never start from scratch.
Ghost Decisions Tracking
Silent objective drift detection that flags code contradicting recorded policies without supersession.
Cryptographic Bounds
Immutable cryptographic audit trails for AI reasoning pathways and execution corridors.
Rationale Corroboration
Real-time decision logic verification against live web telemetry to prevent AI hallucination policies.
Agent Handoff Packets
Role-aware context bundles carrying complete rationale intact across inter-agent boundaries.
Force-Directed Graph UI
Interactive real-time network mapping of decisions, confidence scoring, and causal chains.
Zero-Trust Safety Envelopes
Asimov alignment protocols enforcing hard safety parameters during autonomous operations.
Semantic Citation Search
Instant vector search across citations, documentation graphs, and decision rationale strings.
Multi-Agent Swarm Governance
Centralized control center for monitoring, constraining, and directing multi-agent swarms.
MCP Server Integration
Native Model Context Protocol support for Claude Code, Cursor, Devin, and Zed IDEs.
VSCode Memory Lens
Embedded sidebar viewer providing live inline annotations and reasoning history directly in editor code.
Web Dashboard (Port 8766)
Browser-based command center running locally at port 8766 with real-time telemetry streaming.
Terminal UI & OpenCode CLI
Textual dashboard with interactive slash commands and CLI management.
Silent Drift Warnings
Proactive alerting when agent behavior begins deviating from core project guidelines.
Multi-Source Ingestion
Parallel web, code repository, and paper scanning synthesized into citation-backed knowledge bases.
Zero-Latency State Rollbacks
Instant snapshotting and state rollback to previous clean decision nodes.
Heuristic Adaptation Engine
Dynamic reasoning path adaptation based on newly ingested dataset telemetry.
Role-Aware Security Scoping
Granular permission scoping ensuring agents operate strictly within assigned authorization scopes.
Linux-Native Git Hooks
Seamless hook integration into existing Git workflows, CI/CD pipelines, and commit triggers.
Audit-Ready Compliance Exports
One-click generation of comprehensive safety, alignment, and rationale compliance reports.
Visualize the Rationale
Interactive force-directed graph of decisions. Track causal chains, confidence scoring, and complex relationship mapping in real time.
Core Mechanisms
Advanced telemetry and control systems designed to maintain agent alignment during autonomous execution.
Ghost Decisions
Silent objective drift detection. Flags when code contradicts active decisions without a formal recorded supersession.
Rationale Corroboration
Automatically fact-checks decision logic against live web data to prevent hallucinations from cementing as policy.
Agent Handoff Packets
Role-aware context bundles for multi-agent workflows. Carries rationale intact across inter-agent boundaries.
Start Building with Rationale.
Tropelex is Linux-native, open-source, and integrates directly with your existing Git workflows. Deploy the safety infrastructure today.
View on GitHub