Archived

Nexus
Phase 0

The planning that preceded the product

An ambitious effort to bring Amplifier from CLI to desktop.
What was delivered: not code — a comprehensive blueprint.

The Problem

The CLI ceiling

Amplifier proved that AI-augmented development worked — but only for developers comfortable in a terminal. The question was whether Amplifier could reach a broader audience.

⌨️

Terminal-Only Access

Every interaction required CLI fluency. Non-technical collaborators — PMs, designers, executives — were locked out of the workflow entirely.

👁️

Invisible Work

Sessions, agent activity, context flow — all happening in scrolling text. No way to visualize what Amplifier was actually doing or monitor multiple agents.

🔀

Scaling Friction

Managing multi-agent sessions, switching between projects, and reviewing outputs required juggling terminal windows. The interface didn't scale with the system.

The Vision

Nexus: Amplifier for the desktop

A native desktop application that would surface Amplifier's capabilities through a visual interface — session management, agent monitoring, rich output rendering, and collaborative workflows.

🖥️

Desktop-Native UI

React/Next.js frontend with Electron packaging. Real-time session visualization, multi-panel layouts, and rich markdown/code rendering.

🔌

API-First Backend

FastAPI backend exposing Amplifier's capabilities through REST and WebSocket APIs. Stream agent output in real-time.

👥

Broader Audience

Designed for PMs reviewing AI-generated specs, designers seeing live previews, and executives monitoring team-wide AI activity.

The Core Tension

Distro vs Product

The Nexus planning process surfaced a fundamental question that shapes every platform's evolution.

Distro Model

Amplifier as a composable toolkit. Bundles, modules, recipes — users assemble their own workflows. Power through flexibility. The CLI philosophy.

Strengths: Infinite customization, community-driven, low opinionation, developer-first

vs

Product Model

Nexus as an opinionated application. Curated experience, guided workflows, visual polish. Power through simplicity. The desktop philosophy.

Strengths: Broader reach, lower learning curve, consistent experience, enterprise-ready

Why this tension matters

Every platform faces this fork — Linux chose distro, macOS chose product, VS Code found a middle path. The 153 user stories in the Nexus repository document the team's attempt to navigate this same question for Amplifier.

Deliverables

What Phase 0 actually produced

No code was written. The entire output was planning artifacts — and that was the point. Phase 0 was a deliberate planning exercise.

All artifacts stored in .amplifier-app-docs/ within the repository. Total size: ~850KB of strategy documents.

Velocity

10 days of focused planning

12
Commits
10
Calendar Days
1
Contributor
~40
Planning Files
Jan 28, 2026
First commit. Repository initialized, early vision documents.
Jan 29–31
Competitive analysis, user persona definitions, initial user stories drafted.
Feb 1–4
Architectural planning, API design documents, UX pattern specifications.
Feb 5–6
Synthesis document completed. 153 user stories finalized. 19-week roadmap locked. Last commit.

All commits authored by Mark Licata. A separate repository (amplifier-app-api) contains early backend prototyping by Mark and Alex.

Outcome

Planning complete. Implementation paused.

The repository now sits in the Inactive/ folder. Phase 0 answered the question it set out to answer — what would it take? — and the answer informed a strategic decision.

📐

Scope Was Honest

The 19-week roadmap and 153 user stories revealed the true scale of building a desktop product. This wasn't scope creep — it was scope clarity.

⚖️

Distro Won (For Now)

The core team chose to invest in the composable toolkit — bundles, recipes, foundation — rather than a single opinionated product. The distro model had more leverage.

🗂️

Blueprint Preserved

The planning artifacts remain available. If Amplifier revisits the desktop question, Phase 0 provides a comprehensive starting point — not a blank page.

The lesson

Deliberate planning that leads to a "not yet" decision isn't failure — it's the planning doing its job. Phase 0 prevented months of misallocated implementation effort.

Sources & Methodology

How we verified these claims

All data points in this presentation were gathered from the repository directly. No claims are made beyond what the commit history and file system confirm.

Commands Run

# Repository location ls ~/dev/ANext/Inactive/amplifier-nexus # Commit history git log --oneline --all # → 12 commits total # Date range git log --format="%ai" --reverse # → 2026-01-28 to 2026-02-06 # Contributors git shortlog -sne --all # → Mark Licata (10), mark licta (2) # Code files find . -name "*.py" -o -name "*.ts" # → 0 results # Test files find . -name "*test*" # → 0 results

Key Findings

  • Repository status Inactive/archived
  • Source code files 0
  • Test files 0
  • Planning documents ~40 files
  • Total repo size ~850 KB
  • Directory structure .amplifier-app-docs/ + .git

Data gathered: 2026-02-20. An earlier version of this deck contained fabricated metrics (42 tests, frontend code, etc.) that did not match the repository contents. This version corrects the record.

Amplifier Stories

Honest stories.
Real data.

Every presentation in this series is built from repository evidence —
commit history, file counts, actual code. No embellishments.

More Amplifier Stories