FlyDocs v1.0 ships today. Cloud workspaces, the relay API, the web portal, and full Linear and Jira integration are all live for teams.

This post covers what FlyDocs is, why we built it, and what changes for your team the day you install it.

The problem

Every team we’ve worked with adopted AI coding tools faster than they updated the process around them. Pick your stack: Cursor, Claude Code, Warp, Codex, Windsurf. The code started arriving faster, and everything else fell behind.

The pattern is the same in every codebase:

  • Each developer has their own setup, their own prompt habits, their own idea of what “done” means.
  • The PM tool drifts further from reality every week. Tickets stay open that should be closed, others get marked done without acceptance criteria, and status changes happen in heads instead of in Linear or Jira.
  • Standards live somewhere the agent will never read them: a CONTRIBUTING.md, a Slack thread from six months ago, someone’s head.
  • Onboarding a new developer is a week of “here’s the unwritten rule about the auth layer” sessions.

Faster code generation made all of this worse. The AI is happy to skip the parts of the process that aren’t enforced. So those parts stop happening.

What FlyDocs is

FlyDocs is the workflow layer for AI-assisted engineering. It sits between your developers’ AI coding tools and your team’s process, and it makes the second one non-negotiable.

Skillsnot promptsserver-managedGatesnot guidancenon-bypassable hooksPM Syncalways currentlive to Linear or Jira

Three pieces, working together:

  • Skills, not prompts. Server-managed capabilities that give every AI session the same context, the same standards, and the same procedures. Configured once in the portal, distributed automatically to every repo. Same skill, every developer, every session.
  • Gates, not guidance. Hooks fire on prompts, edits, and status transitions. They format code on save, block status changes without a comment, and stop incomplete work from being marked done. The AI can’t bypass them; they aren’t instructions it might follow, they’re gates it can’t.
  • Live PM sync. Every issue, transition, and comment flows to Linear or Jira automatically through the relay API. Developers stay in the IDE. Leads open their tracker and it reflects what’s actually happening, not what someone remembered to update.

That’s the whole product. Everything else (multi-repo workspaces, persistent session memory, codebase context generation, the workflow lifecycle) serves those three.

What ships in v1.0

  • Cloud workspaces at app.flydocs.ai. Configure your skills, hooks, workflow rules, and provider connections in one place; every developer’s CLI picks up the same config on the next session.
  • Relay API. Provider-agnostic translation between FlyDocs operations and Linear or Jira. Encrypted provider credentials, decrypted only per-request.
  • Multi-repo workspaces. Sibling-repo topology with skills and scripts at the workspace root, per-repo configuration and context.
  • Cross-platform AI tool support. Cursor, Claude Code, Warp, Codex, and Windsurf are first-class. Anything else that reads AGENTS.md works in compatibility mode.
  • Three tiers. Free (local, 1 repo, no account), Solo ($15/mo, 1 workspace, 5 repos), Team ($25/user/mo, 3-20 seats). See the pricing page.

Who it’s for

FlyDocs is built for engineering teams that have already adopted AI coding tools and are starting to feel the gap between code velocity and everything else. Teams that have:

  • Standards that aren’t being enforced consistently
  • A PM tool that doesn’t match reality
  • Multiple AI coding tools across the team and no shared workflow
  • Multi-repo or multi-stack projects where context evaporates between sessions
  • A growing onboarding tax for new developers

It also works for solo developers who want a real framework around their AI session without a team. The free local tier gives you the full methodology with no account.

How to start

The fastest path:

  1. Sign up at app.flydocs.ai and connect Linear or Jira.
  2. Run npm install -g @flydocs/cli && flydocs init in your project.
  3. Open your AI coding tool and run /start-session.

If you want to evaluate without an account first, run flydocs init --tier local. Everything stays on your machine. You can connect a cloud workspace later without losing local work.

~/projects/flydocs-marketing$flydocs init✓Authenticated asmatt@flydocs.ai✓WorkspaceFlyDocs Marketing✓Stack detectedAstro · TypeScript · React✓Generatedflydocs/context/project.md✓Installed12 skills · 7 hooks · 19 commands✓Setup complete in47s→Run /onboard to start your first session.

What’s next

Over the next few weeks we’ll be publishing follow-up posts on the relay, the Linear and Jira setup flow, how skills compare to MCP servers, and what we’ve learned dogfooding FlyDocs on its own codebase. Subscribe to the changelog, follow the roadmap, or join the Discord to follow along.

Welcome to v1.0.