Omni loop research Labs

Posted on Jun 18

Clioloop: An Open-Source AI Agent with Agentic Fusion

#ai #agents #opensource #python

The problem with single-model AI assistants

Most AI assistants give you one model's answer. If it's wrong, you catch it or you don't. If you use a cheap model, quality drops. If you use a frontier model, you pay frontier prices for everything — even a simple file rename.

We wanted something better. So we built Clioloop.

What is Agentic Fusion?

Agentic Fusion puts a whole team of models on one task. When you run /fusion:

Planners (up to 5): Read-only models that research and propose routes in parallel. They figure out the best approach but can't touch your files or run commands.
Main model: Your chosen model does the actual work — full tool access, fully visible. You watch every step. Not a black box.
Reviewers (up to 5): Read-only models that critique the draft. They can see images the main model generated. They check for errors, suggest fixes, flag issues.
Verdict loop: The draft is revised until reviewers approve. The answer you get has already passed independent review.
Fusion: Everything combines into one reviewed, approved answer.

The key insight

The quality comes from synthesis — not from running the same job 5 times. Cheap open models combine into something that rivals a frontier model, at a fraction of the cost.

Safety by construction

Planners and reviewers are read-only at the schema level. They can research and critique, but they can never touch your files or execute commands. Only your main model has tool access, and you watch it work live.

Beyond fusion

Clioloop is also:

Self-improving: Keeps MEMORY.md and USER.md, updated automatically as it learns your preferences and projects
Autonomous: Set a standing goal with /goal and it loops — planning, running tools, checking progress — until done
Everywhere: Terminal, desktop app, web dashboard, Telegram, Slack, Discord, WhatsApp — same session, same memory
Multi-agent Kanban: Break big work into a board of tasks; worker agents pick them up and report back
Tools: File editing, shell, web search & extract, cloud browser, image generation (ComfyUI/Flux), video generation, TTS, any MCP server
Scheduled jobs: Run it on cron for automated workflows
Open-source: Self-host everything. Own your data. No vendor lock-in.

The Omni Loop Portal

One OAuth login gives you access to 300+ models. No API keys to manage. An OpenAI-compatible proxy means you can point any existing tool at it.

The tool gateway handles:

Web search & extract (Firecrawl)
Browser automation (Browser Use)
Image generation (FAL/Flux)
Video generation (Vidu)
Text-to-speech (OpenAI TTS)

All metered, all behind one account.

Install

# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/Clioloop/Clioloop-agent/main/scripts/install.sh | bash

# Windows (PowerShell)
iex (irm https://raw.githubusercontent.com/Clioloop/Clioloop-agent/main/scripts/install.ps1)

Then:

clio setup    # pick Omni Loop Portal, approve in browser
clio          # start looping

Architecture overview

The repo is a monorepo:

Python core (clio CLI + agent loop in clio_cli/ and agent/)
TypeScript surfaces (Electron desktop, React-Ink TUI, React web dashboard)
Omni Loop Portal (Next.js 15 + SQLite + Stripe)

Every provider is a ProviderProfile with lazy discovery. The managed provider is the Omni Loop Portal (OAuth device-code, no API key). Model catalog + switching is cached locally.

The fusion engine lives in agent/fusion_engine.py. Slash commands have a single source of truth in clio_cli/commands.py.

DEV Community