Intelligence layer · Built above TX-1 Terminal
Intelligence before
the screenshot.
SS-1 monitors your competitive landscape continuously and delivers a structured intelligence brief — before the LinkedIn post arrives and the reactive sprint begins.
5
LangGraph agent nodes
0
manual research steps
100%
HITL approval required
git
audit trail for every brief
The problem
Product teams are reactive by default. Not because they want to be — because competitive intelligence is a manual, periodic, expensive process.
Tuesday
Competitor ships
A competitor releases a feature. No one on your team knows yet. The changelog is live.
Wednesday
LinkedIn screenshot
Someone in the industry posts about it. Your PM sees it and screenshots it in Slack.
Thursday
Reactive sprint
A feature request lands in your backlog with no strategic context, no user impact analysis, no assessment of whether responding is even the right move.
The status quo — in 15 seconds.
What SS-1 does
Six capabilities, one terminal.
Continuous manifest monitoring
Your Project Manifest is a git-backed SSOT: vision, competitors, archetypes, priorities. SS-1 monitors the external world against it — continuously.
LangGraph intelligence pipeline
Five agent nodes: Monitor → Analyst → Significance Gate → Propose → HITL. Haiku for routing and scoring. Sonnet for analysis and brief generation.
Significance scoring
Not every competitor move deserves a response. Every signal is scored 0–100. Below 40 is discarded silently. Above 40 surfaces a card for your review.
Structured intelligence briefs
Not a news summary. A structured brief: what changed, strategic implication, proposed manifest update, and three strategic response options.
HITL approval gate
No manifest update without explicit user approval. Every agent-proposed change is previewed as a diff before it touches the file.
Git-versioned audit trail
Every approved brief creates a git commit. The manifest's evolution is an auditable record of how your competitive landscape changed — and how your thinking responded.
Live walkthrough
See it in action.
Watch a complete competitive sync — from cold terminal to approved manifest update. One command. The full five-node pipeline. Human approval. Git commit. 26 seconds.
What the walkthrough covers
Key design decisions
Six choices that shaped the architecture.
01
Why a manifest file instead of a database?
A markdown file under git gives you a diff, a commit history, and a human-readable record. A database gives you none of those. The manifest is designed to be read by a human, not just queried.
02
Why HITL approval on every manifest write?
An agent that updates your strategy without asking is not an intelligence tool. It's noise. Every proposed change must pass through a diff preview and explicit approval before it touches the file.
About the builder
Anthony Key
Product designer and builder working at the intersection of AI systems and product strategy. SS-1 is the intelligence layer above TX-1 Terminal — together they represent a complete thesis about where enterprise AI tooling is heading: systems that don't just respond to what you ask, but anticipate what you need to know.
Every architectural decision in SS-1 is documented and defensible. This is not a side project. It is a portfolio artifact demonstrating AI product architecture, agentic systems design, and the kind of senior product thinking that connects technical choices to strategic outcomes.
Stack
Tauri v2
Desktop shell
React 19
Webview UI
LangGraph
Agent orchestration
FastAPI
Python sidecar
Claude API
Haiku + Sonnet
Firecrawl
OSINT crawler
SQLite
Local persistence
GitPython
Manifest versioning
TX-1 · SS-1
The execution layer. The intelligence layer.
TX-1 acts on problems inside your systems. SS-1 monitors for problems emerging outside them. The manifest schema connects them. Together they represent a complete picture of enterprise AI tooling — reactive and proactive, unified.