Getting Started

Set up Fenix and connect your AI agent in under 2 minutes.

Fenix is an AI-native project management workspace built for teams that use AI agents. Every agent you work with gets direct, real-time access to your tasks, memory, docs, and rules through a native MCP bridge.

What you get

Feature	What it does
MCP Bridge	Any MCP-compatible agent reads your tasks, memories, and docs in real time
Semantic Memory	Context persists across AI sessions — agents remember your decisions
Rules & Skills	Shareable AI instructions and workflows, exportable to any agent
Kanban + Sprints	Full project tracking with boards, sprints, and burndown
Docs & Wiki	Built-in knowledge base linked to work items
CLI + TUI	Native terminal interface with vim motions, no browser needed

Choose your path

Option A — Web app (fastest)

Go to app.fenix.devshire.app
Sign up for a free account
Create your workspace and first team
Start adding work items

Free for teams up to 5. No credit card required.

Option B — CLI + MCP (AI-native setup)

# Install via pipx (recommended)
pipx install fenix-mcp

# Or via pip
pip install fenix-mcp

Then connect to your agent — see the MCP installation guide.

Option C — npx one-liner

npx fenix@latest init

Scaffolds a .fenix/ workspace directory, creates a starter board, and prints the MCP config snippet for your agent.

Core concepts

Work item hierarchy

Epic
└── Feature
    └── Story
        └── Task / Bug

Memory

Structured context stored with semantic embeddings. When you save a decision, your agents can retrieve it in future sessions and avoid re-asking or contradicting past choices.

MCP Bridge

A local server exposing 68+ actions over the Model Context Protocol. Connect once and every MCP-compatible agent gets real-time read/write access to your workspace.

Quick example

Once the MCP bridge is running:

you: "what's blocking sprint 14?"

agent → reading sprint:S-14, memory:decisions…
  2 blockers · PROJ-248 auth redirect, PROJ-251 cli autocomplete
  both tied to decision "keep SSO on same domain"
→ answered from real project data · no hallucination