Docs

Local Installation

Install and configure FlyDocs for local-only use. No account required.

Local installation uses the same CLI as cloud setup. The difference is that you skip the account connection step. FlyDocs detects the absence of an API key and defaults to the local tier.

Install the CLI

Terminal 
npm install -g @flydocs/cli

Requires Node.js 18.17+ and Python 3.8+.

Run It in Your Project

Terminal 
cd your-project
flydocs

The installer walks you through setup. When it asks about connecting an account, skip that step. FlyDocs detects no API key and configures the local tier automatically.

What happens during install

  • Stack detection: scans your project for frameworks, languages, and tooling
  • Skill generation: recommends and installs community skills matching your stack
  • Config creation: generates .flydocs/config.json with "tier": "local"
  • Hook installation: sets up all 7 hooks for context injection, auto-formatting, validation, and enforcement
  • Context generation: creates project context, knowledge base, and agent definitions

Local Configuration

The config file tells FlyDocs to use the local backend for all operations.

.flydocs/config.json 
{
  "tier": "local",
  "provider": {
    "type": "local"
  },
  "workspace": {
    "activeProjects": []
  },
  "statusMapping": {
    "BACKLOG": "Backlog",
    "READY": "Ready",
    "IMPLEMENTING": "In Progress",
    "BLOCKED": "Blocked",
    "REVIEW": "Review",
    "TESTING": "Testing",
    "COMPLETE": "Done"
  }
}

Status mapping still exists in local mode. It maps FlyDocs lifecycle stages to display names used in local issue files. You can customize these names if you prefer different terminology.

Your Project After Install

The installer creates these directories:

Project Root 
flydocs/                       # Project context and knowledge base
├── context/
│   └── project.md             # Stack, scope, and priorities
├── knowledge/                 # Session decisions and learnings
└── issues/                    # Local issue storage (local tier only)

.flydocs/                      # Configuration, hooks, and scripts
├── config.json                # Tier, provider, status mapping
├── hooks/                     # Deterministic hook scripts
└── _local/                    # Local backend implementation
    └── file_store.py

.claude/                       # Claude Code integration
├── skills/                    # Workflow and stack skills
├── commands/                  # Slash commands
├── settings.json              # Hook configuration
└── agents/                    # Sub-agent definitions

.cursor/                       # Cursor integration (if applicable)
AGENTS.md                      # Cross-platform manifest

Everything Works the Same

With local mode configured, every feature works identically to the cloud tier except issue storage:

  • Hooks fire: all 7 hooks across 5 event types are active
  • Skills load: workflow methodology, stack patterns, and coding conventions
  • Slash commands work: /capture, /implement, /review, and all others
  • Enforcement is active: stop gates block incomplete work, transitions require comments
  • Context injection runs: project context and session state are injected on every prompt

The only visible difference: when you create an issue, it is stored as a local markdown file instead of syncing with Linear or Jira.

Need Help?

  • Join the Discord for support, feature requests, and community discussion.
  • See Installation for the full install reference including prerequisites.

Next