Clero

Inspiration: Building a solo project, I realized there’s no tool that can lead projects with an AI team. Paperclip was a part of my inspiration, but I was thinking about creating an easy-to-use experience for middle-range users who want to lead projects with AI. Codex and Cursor were another inspiration. I wanted a minimalistic tool that would empower users to run projects that aren’t just about code, but also about business, communication, and more.

What it does: It recommends a team of AI agents for every project, each project led by a project manager AI agent. Each agent has its own model, allowing for the benefits of a model-agnostic solution. Agents can conduct product research, make changes, communicate with third parties, connect to tools like Cursor, Codex, Antigravity, and Claude CLI, and prompt them just like a real person. However, since they lead the project, they will have the full context. Specialist agents work across Slack, Telegram, code, and documentation. You’ll witness the project board transition from “to do” to “done.”

We built our project using Python/Django/Temporal/Celery and Nuxt/TypeScript/Tauri for MacOS. I’ve been using Codex since the project’s inception, and GPT-5.6 Sol unlocked the ability to complete the most challenging tasks. In the final hours, I deployed my project to AWS.

Challenges: Multi-agent coordination (race conditions, dependencies), model reliability across providers, solo-founder prioritization with paying customers on the line, macOS app (never had any experience)

Accomplishments: Model-agnostic architecture that works. ~10 paying customers at $190 MRR. Agent delegation that actually ships. Zero-downtime model failover. Built alone, shipped globally.

We’ve learned that it’s impossible to target everyone and be competitive simultaneously. You must prioritize features and consider the long-term consequences of your decisions.

What’s next? More project-scoped features and a project development cycles feature where agents can run autonomously, handling everything themselves. Additionally, there’ll be improved integration with codebases and coding platforms like Codex.

Built With

• amazon-web-services
• celery
• django
• nuxt
• python
• tauri
• temporal
• typescript