Compare All Features

Multi-Agent Code Editing (Verdict Code)

AI-powered code editing with multi-file support via Verdict Code

Prime Directives Execution

Submit project goals and get executed by PAS agent swarm

PAS (Project Agentic System)

Multi-tier agent orchestration (Architect, Directors, Managers)

Blueprint (Project Estimation)

Dry-run project planning with cost/timeline/energy estimates

PLMS (Project Lifecycle Mgmt)

Multi-run project tracking, receipts, KPI validation, Bayesian calibration

Local Model Support (Ollama/LM Studio)

Use local open-source models (Llama, Qwen, etc.)

BYOK (Bring Your Own Keys)

Use your own Anthropic/OpenAI/Gemini/Grok API keys

Verdict Cloud Proxy Mode

Verdict-owned vendor keys (Anthropic, OpenAI, Gemini, Grok)

LAN Model Endpoints

Connect to OpenAI-compatible endpoints on local network

Auto Model Routing (4 modes)

auto_local, auto_byok, auto_cloud, auto_general + quality tiers (Low/Medium/High)

Model Fabric Caching

TTL-based in-memory cache for model discovery (60s TTL, 110x faster)

RAG Code Retrieval (LightRAG)

Graph-based code index with query-first semantic retrieval

Verdict Web (Publish/Host/Share)

One-click publish static web projects to live URLs with QR codes and share links

Share Links (Verdict Web)

Generate expiring share links and QR codes for published projects

Semantic Chunking

ML-based code/doc chunking for retrieval

Password Protection (Verdict Web)

Add password protection to shared project links

File Manager

Web-based file browser and editor

Git Integration

Auto-commit, diff, status tracking

Sandbox Execution

Isolated code execution environment

Code Linting (Ruff)

Automated code quality checks

Telemetry Service (CPU/MEM/GPU)

Real-time resource usage metrics

Process Tree Walking

Auto-discover descendant processes of agents

JobCard Store (Persistence)

Durable state storage for all PAS JobCards

Heartbeat Monitor

Service health tracking and failover

Service Registry

Dynamic service discovery via registry.db

Comms Logging

Parent-child chat logging with LLM metadata

Personal Billing Account

Individual user credit management

Organization Accounts (Teams)

Shared billing pool with role-based access

Usage Analytics Dashboard

System-wide usage and revenue metrics

Billing Export (CSV)

Download transaction history

Cost Attribution

Token usage metering by model/user/org

Ledger Events Tracking

Append-only audit trail of all billing transactions

Hard Cap Enforcement

Spending limits per account

Credit Adjustment (Admin)

Manual credit grants/deductions for support

RBAC (Role-Based Access Control)

7 roles x 10 GUI areas with fine-grained permissions

Permission Overrides

Grant specific permissions outside role defaults

super_admin_lock

Protects platform owner role from accidental removal

API Keys Management

Generate/revoke API tokens with scopes

Key Storage (Encrypted)

SQLite key vault for BYOK credentials

Cloud BYOK Vault

Optional zero-knowledge encryption for key sync

OAuth Device Token

Installed app authentication (device-code flow)

Secrets Redaction (Logs)

Auto-remove API keys, tokens, credentials from output

Filesystem Allowlist

YAML config to whitelist/blacklist file access

Command Allowlist

YAML config to block destructive git/rm operations

Audit Logging (RBAC)

Complete trail of role changes and permission denials

Organization Management

Create orgs, manage members, set billing policies

Multi-Org Support

User can belong to multiple organizations

Team Chat (Parent-Child)

Hierarchical message threading with agents

Receipt Generation

JSON receipts for every PAS execution

Checkpoint Storage

JobCard state snapshots for resume/rollback

Web HMI

Browser-based dashboard for PAS control

Tree View (D3.js)

Hierarchical agent/task visualization

Agent Chat Interface

Send messages to agents with tool context

Sequencer Panel

Timeline of execution steps with audio notifications

Job Card Dashboard

View active/completed jobs and statuses

Telemetry Widget

Real-time CPU/MEM/GPU metrics display

Settings Panel

Configure models, feature flags, provider keys

Auto Model Picker

Dynamic model selection with filtering

Chat Filters

Filter chat messages by type (text/tool/error/result)

Voice STT (Speech-to-Text)

Convert audio input to text

Voice TTS (Text-to-Speech)

Convert agent responses to audio

Audio Notifications

Completion/error alerts with multi-lane synthesis

Model Compatibility Testing

Auto-detect tool-calling and code editing support

SLO Validation Endpoint

Check service SLOs and health

Performance Benchmarking

Model latency and throughput metrics

SAM (Shared Agent Memory)

Shared context across agents: Brief injection, decision ledger, concurrency control

Remote Control (Monitor)

Monitor Verdict status and agent progress from iPhone, iPad, Apple Watch, and Android devices

Remote Control (Monitor and Control)

Full remote control of Verdict from iPhone, iPad, Apple Watch, and Android devices - monitor status and send commands

Holoview (Holo Layer)

3D visualization of agent architecture and data flows

Vendor Error Mapping

Normalize errors across Anthropic/OpenAI/Gemini/Grok

Speculative Execution

Early token generation for latency reduction

Relay Streaming

SSE passthrough of vendor streams

Model Resilience

Auto-failover to backup models on error

SSO / SAML

Single Sign-On via SAML 2.0 with IdP integration

Dedicated Infrastructure

Single-tenant deployment option with isolated networking and keys

Custom SLO / SLA

Contracted uptime and support response guarantees

Custom Data Retention

Configurable retention policies for logs, receipts, and artifacts

Audit Log Export

Export audit logs to external systems (SIEM) and scheduled archives