Frequently asked questions.

Storybloq is project memory and workflow for AI coding. It keeps your plans, stories, decisions, and review records with your project, helping Claude Code and Codex carry the work across sessions.

People building ongoing projects with Claude Code or Codex, from a first prototype to an established codebase. Start with stories and handovers. Add review and coordination as the work grows.

No. Your coding client writes and runs code. Storybloq gives it a shared project record and structured workflows. You keep your editor, Git repository, and AI provider.

Node.js 20 or later and a local project, plus Claude Code or Codex for the guided Story workflow. The Mac dashboard is optional. AI subscriptions, API usage, and review backends may have separate costs.

Run npm install -g @storybloq/storybloq@latest, then storybloq setup --client all. Open your project in Claude Code or Codex and invoke /story or $story to load an existing ledger or start guided setup. Re-run the install and setup commands to upgrade.

A written record of what changed, why decisions were made, what is blocked, and what comes next. Ask your agent to leave a handover before ending a session. The next session can load it through /story or $story; continuity depends on keeping that record current.

Yes. The CLI, MCP server, and Story skill work without the Mac app. The optional dashboard gives you a visual view of stories, issues, handovers, and session state.

Yes. Stories, handovers, lessons, and review records are useful in everyday collaborative sessions. You choose when to add autonomous execution or more agents.

It guides a supported coding client through scoped work: planning, plan review, implementation, tests, code review, and completion. It records progress for recovery. It depends on the client being able to run and use the required tools; it is not a promise of unattended completion or automatic wake-up.

No. Storybloq does not certify that generated code is correct. It preserves plans, review findings, verdicts, and project history so you can inspect the evidence. Independent AI review can still miss bugs. Tests, CI, and release checks remain necessary.

Both have setup for the Story skill and MCP project tools. Hooks, agent dispatch, messaging, and completion notifications depend on the client and tools available in that session. Shared project files do not imply identical execution capabilities.

Claude Code and Codex have maintained setup flows. Other tools can read the JSON and Markdown files, and compatible clients can use the MCP server. File portability does not mean every client has the same skill, hooks, or orchestration support.

Yes. A shared wrapper ledger can describe several repos, and federation can coordinate repositories that keep their own ledgers. These are different layouts; see the multi-repo guide and federation documentation before choosing one.

Yes. CLAUDE.md and AGENTS.md remain useful for project rules. Storybloq adds changing stories, decisions, and handovers. You can keep another issue tracker, but automatic synchronization requires a separate integration.

The CLI and Mac app are free to use. The public CLI, MCP, and skill source is source-available under PolyForm Shield 1.0.0, not an OSI open-source license. The license restricts competing uses; read the LICENSE and NOTICE in the public repository. Your AI provider and optional services may charge separately.

Core project records are JSON and Markdown files in your local .story/ directory. You decide what to commit and push to Git. Connected AI clients can send the context they read to their providers. Optional companion access and feedback also use network services; local storage does not mean every workflow stays on your device.

The privacy policy separates local project files, AI-provider connections, paired companion devices and Apple CloudKit, submitted feedback, optional Mac diagnostics, and website analytics. Mac analytics and crash reporting are off by default. The website uses Vercel Web Analytics separately from the Mac preference.