What is Claude Code in Glueprint?
Claude Code is a first-class native adapter in Glueprint — not a wrapper script, not a re-implementation, and not routed through the generic ACP harness. Four things make it feel native.
Everything below works with the Claude Code you have today — no migration, no new account for the model, no lock-in. Uninstall Glueprint and your terminal workflow is exactly as it was.
Set up Claude Code on a host.
Installation is host-specific: repeat these steps anywhere this agent should execute. Portal and mobile can control the session afterward, but the process remains on that host.
Install and authenticate Claude Code
Install the official package on every host where sessions should run, then complete Claude Code's own login flow.
Detect the CLI in Glueprint
Open Settings -> Agents -> Claude Code. Leave CLI Path on Auto-detect, choose Detect, and confirm the resolved path and version.
Start a project session
Open a project or worktree, create a session, choose Claude Code, and send a small prompt. Glueprint records the Claude session ID for later resume.
npm install -g @anthropic-ai/claude-code
claude --versionHow the connection works.
Glueprint adds orchestration and shared surfaces without taking ownership of the coding agent or its provider account.
Glueprint finds the claude CLI you already installed — on PATH or at a custom path you point it to — probes the version, and spawns it. Your login, your subscription, your settings, your project's CLAUDE.md: all of it comes along, because it's the same binary you run in the terminal.
The adapter runs Claude Code in stream-JSON mode with partial messages enabled, and parses stdout into a typed event timeline: prompts, reasoning, tool calls with arguments and results, diffs. That's what makes sessions watchable from the portal or a phone instead of a log file.
When Claude Code wants to write a file, run a command, or call an MCP tool, the request routes through Glueprint's permission prompt handler and lands as an inline approval card — on the desktop, in the browser, or on your phone. Approve, deny, or auto-allow per tool.
Glueprint stores the Claude session id with every work item and resumes it with --resume. Restart the app, reboot the host, switch from desktop to portal mid-task: the conversation picks up in the same context window, not a fresh one.
Capabilities and discovery.
The status labels separate guaranteed integration behavior from features owned or announced by the installed agent.
| Capability | Status | Behavior |
|---|---|---|
| Structured streaming | Supported | Native stream-JSON parsing with partial messages. |
| Session resume and fork | Native | Persisted session IDs; resume and fork use Claude's own controls. |
| Permission approvals | Supported | MCP permission flow becomes Glueprint approval cards. |
| Images | Supported | Images can be attached to supported Claude sessions. |
| Models, modes, commands | Supported | Discovered from the installed CLI and changeable while supported. |
| Reasoning effort | Native | Low through max can be selected per session. |
| Explicit interrupt | Supported | Active turns can be interrupted from any connected surface. |
Authentication, ownership, and limits.
Glueprint is the environment around the agent. Credentials, model access, and vendor-specific behavior stay where their owner designed them.
Watch every prompt, tool call, and diff stream in real time — desktop, web portal, iOS and Android. The session runs on your machine; the view follows you.
Permission prompts arrive as approval cards wherever you are. Kick off a refactor at your desk, approve the file writes from the couch.
Run several Claude Code sessions side by side, each in its own git worktree. Built-in worktree management, diffs, and a three-pane conflict resolver keep them from stepping on each other.
Tile Claude Code sessions from your laptop, your desktop, and a homelab box into one Command Center mosaic. Every tile carries a host.
Set thinking effort per session — low through max — instead of one global default. Cheap loops stay cheap; hard problems get full depth.
Promote Claude Code to a persistent assistant: persona, constitution, long-term memory, and a daily journal on disk — carried across wakes by resuming the same session.
Wake Claude Code on cron: triage the inbox every 30 minutes, sweep PRs at 9am, run the nightly cleanup. Quiet hours and watermarks prevent re-fires.
API value, messages, sessions, and output tokens computed from the session logs Claude Code already writes — filterable by project, host, or assistant.
Common questions
Does Glueprint replace Claude Code?
No. Glueprint is an Agentic Development Environment that hosts Claude Code. It spawns the official CLI you already installed and adds the surfaces around it: live session streaming, inline approvals, parallel worktrees, scheduling, memory, and remote control.
Do I need an API key or a separate Claude subscription?
No. Glueprint doesn't resell model access. Sessions run through the claude CLI's own login, so your existing Claude subscription or API setup is exactly what gets used.
Will it use my CLAUDE.md, skills, and MCP servers?
Yes. Sessions launch from your project directory, so CLAUDE.md, skills, and your configuration load exactly as they do in the terminal. Assistant sessions additionally get Glueprint's own MCP tools for memory, journaling, and tasks.
Can I run Claude Code from my phone?
Sessions always run on your machines — a laptop, a desktop, or a server with the headless daemon. The iOS and Android app streams them live, lets you send follow-ups, and handles permission approvals from anywhere.
How many Claude Code sessions can I run at once?
As many as your hardware handles — Glueprint doesn't cap parallel sessions. Pair each with its own git worktree and tile them in the Command Center; paid plans add more hosts to spread the load.
Is my session content private?
Yes. Remote viewing routes AES-256-GCM-encrypted envelopes through the relay, which stores a wrapped key it cannot open — decryption happens on your devices. On the free local tier, session content never leaves your machine at all. See security.