March 2026
v0.7.0 · Current
Tier 3: AI-Autonomous Conversation Engine
New conversational intake mode — the AI drives a guided dialogue to collect all required fields, proposes a decision, and waits for human approval. Multi-turn sessions with per-domain question schemas, conditional logic, auto-fill computed fields, and full audit trail of every exchange. Three-tier architecture: AI-Enhanced, AI-Interpreted, AI-Autonomous.
Conversation Engine
Tier 3 Chat UI
Question Schemas
Decision Proposals
Conditional Questions
539+ Tests
March 2026
v0.6.0 · Shipped
Tier 2: AI-Interpreted Document Intake
Paste unstructured documents — the AI extracts structured fields using domain-specific extraction schemas. IntakeProcessor maps free text into decision context with sample documents for each domain. LLM-powered extraction with fallback to keyword-based parsing.
Document Intake
Extraction Schemas
Sample Documents
Tier 2 UI
February 2026
v0.5.0 · Shipped
Real AI Integration & LLM Hot-Swap
Integrated OpenAI and Ollama as real LLM backends alongside MockAI. Hot-swap between backends at runtime without losing audit state. Model-agnostic architecture with LLMInterface protocol. API key management with server-side persistence.
OpenAI Integration
Ollama Support
LLM Hot-Swap
Model-Agnostic
AI Settings UI
February 2026
v0.4.0 · Shipped
Audit Tamper Detection & Policy Config
SHA-256 hash chain verification with tamper simulation and one-click recovery. Live policy configuration UI for tuning domain thresholds in real time. PDF audit export for compliance documentation.
Tamper Detection
Hash Chain Recovery
Policy Config UI
PDF Export
API Documentation
Pricing Page
January 2026
v0.3.0 · Shipped
What-If Analysis & Decision Flowchart
Counterfactual "What-If" analysis lets users tweak inputs and see how decisions change. Interactive Mermaid.js decision path flowcharts visualize the entire 7-stage pipeline for every decision.
What-If Engine
Counterfactual Analysis
Mermaid.js Flowcharts
Self-hosted Visualization
December 2025
v0.2.0 · Shipped
Multi-Domain & Explainability Dashboard
Expanded beyond health to support Insurance and Finance domains. Built a full explainability dashboard with real-time confidence scoring, evidence breakdown, and rule compliance visualization.
3 Domains
Insurance
Finance
Explainability Dashboard
Confidence Scoring
November 2025
v0.1.0 · Shipped
Core Decision Engine
Initial release of the 7-stage auditable decision pipeline: Risk Assessment → Evidence Collection → Rules Engine → Reasoning → Confidence Scoring → Decision Gate → Audit Trail. Zero external dependencies.
7-Stage Pipeline
Health Domain
Audit Trail
Zero Dependencies
Pure Python