Plumichat — Data skill for Claude Code
Self-hosted web app for driving Claude Code from your phone.
How to install Plumichat
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open THRAUR/plumichat and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Plumichat does
Self-hosted web app for driving Claude Code from your phone. Streaming turns, permission prompts as taps, a real terminal, and two-layer agent sandboxing. Node 22, no build step, no framework, no database.
Alternatives in Data
- Claude Code Usage Monitor — by Maciek-roboblog - A real-time terminal-based tool for monitoring Claude Code token usage 7k ★
- 06 Ink React Terminal UI — Prompt 06: Verify and Fix the Ink/React Terminal UI Pipeline 2.3k ★
- Azure Event Hub .net — High-throughput event streaming 1.8k ★
README
PlumiChat
**Drive [Claude Code](https://claude.com/claude-code) from your phone.**
A self-hosted web app that puts an agentic coding session — streaming output, tool approvals, a real terminal, file browsing — on any device you own, with the work still happening on your machine.
Runs on Linux, macOS and Windows. No build step, no framework, no database.
[Install](docs/INSTALL.md) · [Usage](docs/USAGE.md) · [Architecture](docs/ARCHITECTURE.md) · [Security](docs/SECURITY.md)
A turn in flight — tool calls, rendered output, and the answer still streaming.

Claude Code's permission prompts, as taps. The full command is shown before you approve it.
[!IMPORTANT] **PlumiChat gives a language model a shell on the machine it runs on.** That is the point of it, and it is why it binds to `127.0.0.1` by default and refuses to start on a public interface with no authentication configured. Read [SECURITY.md](docs/SECURITY.md) before exposing it to anything.
Why
Claude Code is a terminal program. Terminals are not usable on a phone: no arrow keys, no scrollback worth reading, no way to approve a permission prompt one-handed on a train.
PlumiChat is that session as a web app. Not a wrapper that shells out to the CLI and prints the output — it drives the Agent SDK directly, so tool calls, permission prompts, questions, background agents and session history are all first-class.
The design constraint throughout: **the primary client is an iPhone home-screen web app on a bad connection.** That single as
Related Skills
Cubby
Self-hosted app hosting for you and your friends' AI coding agents. Ask Claude Code, Codex or Cursor for an ap
Gba Build Develop
/gba-build-develop — Self-Driving GBA Game Development Pipeline
Self Driving Explain Limits
Explain the data, model, optimization, and implementation limits of a run
Self Driving Run
Run a new IPS-governed Strategic Asset Allocation pipeline
Idotmatrix
Turn your iDotMatrix into a live desktop companion — the all-in-one app for your iDotMatrix: 90+ apps, 22 mult
Brightdata Scraper Studio Skill
Claude Code skill for building, running and self-healing Bright Data Scraper Studio web scrapers from the term
Related Agents
ZeroClaw Android
Run AI agents 24/7 on your Android phone. Native Rust core, 25+ providers (OpenAI, Claude, Gemini, Groq, DeepS
Integration Runner
End-of-build closer. Runs the project's real test/validate pipeline and iterates until it is genuinely green,
Goal Relay Agent
Use this agent as the final stage of a completed goal (a pwn-box run, or any bounded security-research task) t