Skip to content

Founding Cohort Demo

Jamaal Linton

Founding Cohort

J's ZeroClaw Lab

A documented zero-to-deployed journey from first server access to a delegation-capable AI agent system running on a low-cost VPS.

Project Snapshot

Builder

Jamaal Linton

Sales Engineer | Datasite

Lead Agent

Nova

A YouTube Intelligence Officer that reviews video links and returns verdict-first briefs through Telegram.

Infrastructure

$6/Month VPS

The live site frames the build as a practical deployment story on low-cost infrastructure rather than a purely conceptual demo.

What The Project Is

The homepage opens with a clear transformation:

Four weeks earlier, Jamaal had never SSH'd into a server. The live site now documents a delegation-capable AI agent system running from a low-cost VPS.

That makes the project valuable for two reasons:

  • it is a working agent build
  • it is also a full record of how the build happened

The site describes itself as a personal knowledge base from the Super Individual Founding Workshop, documenting every step, blocker, and breakthrough from zero to deployment.

The Journey

The published page breaks the build into a four-week arc:

  • Week 1: bought a Hetzner VPS, installed ZeroClaw, connected Telegram, sent the first command
  • Week 2: tightened permissions, encrypted API keys, set hard spending limits
  • Week 3: built Nova and learned delegation
  • Week 4: published the documentation site

That is a strong pattern for a cohort page because it shows operational progress, not just a polished endpoint.

Meet Nova

Nova is framed on the live site as Jamaal's first sub-agent.

Its role is specific:

  • it acts as a YouTube Intelligence Officer
  • it lives in a sandboxed workspace on the server
  • Jamaal sends a video link through Telegram
  • Nova returns a verdict-first brief

The response format is designed around practical decision-making:

  • HIGH ROI or LOW ROI
  • key findings
  • actionable takeaways

That is the right kind of narrow scope for an early agent system. The value is obvious and the workflow is easy to understand.

Blockers And Fixes

One of the strongest parts of the site is the troubleshooting honesty. The live page lists concrete failures and the fix applied:

  • YouTube authentication broke, so the build moved from a browser-cookie package to the Apify YouTube MCP server
  • OpenRouter credits ran out mid-task, which forced better management of max_tokens and model routing
  • tool registration mismatched after migration, which required updating Nova's allowed tools to match the new Apify server names

This is useful evidence that the project is not a surface-level demo. It shows real operator learning.

Documentation Pattern

The live site also exposes a broader documentation structure:

  • architecture sections for Identity, Soul, Skills, User, and Style
  • configuration reference
  • build journal
  • showcase of best outputs

That means the project is being treated as a maintainable system rather than a one-off experiment.

Visit the Demo

Closing Thought

Jamaal's project is a strong founding cohort example because it demonstrates operational progression clearly: learn the infrastructure, secure the system, delegate to a focused agent, and publish the record of how it all worked.