Twenty-four capabilities built around one principle: governed multi-agent orchestration with evidence at every stage.
🔄
Multi-Agent Pipeline
Commander → Coder → Reviewer flow with 10 structured stages. Full context forwarding between every model.
🛡️
Approval Gates
Human governance at every critical stage. Your AI team proposes — you approve, reject, or redirect.
💾
100% Local Storage
Chat history, configurations, and memory stored entirely on your machine. Nothing leaves your device.
☁️
Hybrid Cloud + Local
Mix Claude, GPT, Gemini, Mistral, and Grok with 24+ local Ollama models on the same team. Your choice, per role.
🔥
Warm State System
Pre-load models before you need them. Visual readiness indicators — Cold, Warming, Warm — for every role.
🔍
Web Search Tools
Built-in search via SearXNG and LibreX backends. Your AI team researches the web without leaving the pipeline.
📊
Evidence Cycles
Every decision documented with STATUS_BLOCKs. Proof of work, not promises. The Automated Outcome you can verify.
🧠
10-Stage Pipeline
Handoff → Tandem Review → Map → Verify → Confirm → Build → Quality Loop → Test → Deploy Gate → Ship.
⚡
Team Presets
App Builder, Doc Writer, Website Refactor, and more. One click configures your team and warms the models.
🧪
Thinking Panel
See your AI's reasoning in real time. Full trace visibility into how each model arrives at its decisions.
👥
19 Specialized Roles
Pre-built roles across 7 categories — Command, Design, Build, Quality, Ops, Knowledge, Media — each with a defined abilities matrix.
📶
Skill Levels
Beginner Guided through Advanced. The UI adapts automatically — simpler controls for new users, full power for experts.
🎛️
Simple + Advanced
Pick a preset and go, or take full manual control of all 19 roles, model assignments, and pipeline stages.
🚧
Query Guardrails
Dangerous patterns blocked before they reach your models. Built-in safety layer protects your pipeline and your data.
📡
Real-Time Notices
Live event stream showing pipeline activity, model responses, and system status as it happens.
📦
Import / Export
Chat history as JSON — portable, yours, and never locked in. Export everything, import anywhere.
🔘
Capability Dots
Tools, vision, embed, thinking — see what each model can do with visual capability indicators on every role card.
📏
Storage Meter
Always know your local usage. Auto-prune settings keep your storage clean without losing important conversations.
🌙
Dark Mode Native
Built dark-first, not bolted on. Every panel, card, and control designed for extended-session comfort.
💬
Multi-Conversation
Create, rename, delete, and switch between chat sessions seamlessly. Each conversation keeps its own history and context.
🔀
Sequential & Parallel
Run pipeline stages in order or simultaneously — your choice per mission. Match the workflow to the task.
🔗
Context Forwarding
Full conversation history passes between every stage and every model. No information lost. No copy-paste.
📒
Gate Ledger
Every approval decision recorded with timestamp, role, and outcome. Full audit trail across all 10 pipeline stages.
🌐
Tandem Research
Commander and Coordinator perform real web searches during planning. Two-pass discovery — questions, then evidence.
