shawntabrizi

Polkadot Chat Agents — AI skill for Claude Code

AI community

Run AI agents in the Polkadot app.

How to install Polkadot Chat Agents

This entry records only its repository, not the path inside it, so there is no exact command to give. Open shawntabrizi/polkadot-chat-agents and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Polkadot Chat Agents does

Run AI agents in the Polkadot app. E2E-encrypted and private.

Alternatives in AI

Open Dots — Open-source, self-hosted AI agent workspace and alternative to OpenAI Dots, Meta Muse, Grok Bot, Instinct, Man 5.5k ★
Deepreasoning — A high-performance LLM inference API and Chat UI that integrates DeepSeek R1's CoT reasoning traces with Anthr 5.4k ★
Chat Ollama — ChatOllama is an open source agentic app for running AI agents across local and hosted models 3.5k ★

README

*Run your own AI agents and talk to them from the Polkadot app — end-to-end encrypted, with no chat server.*

[![npm](https://img.shields.io/npm/v/polkadot-chat-agents)](https://www.npmjs.com/package/polkadot-chat-agents) [![Documentation](https://img.shields.io/badge/docs-website-e6007a)](https://www.shawntabrizi.com/polkadot-chat-agents/) [![License: MIT](https://img.shields.io/badge/license-MIT-blue)](LICENSE)

Polkadot app chat has no central server: messages travel over the Statement Store, a decentralized store-and-forward layer, and every conversation is end-to-end encrypted. That means a chat bot is just a process with an outbound connection to a public RPC node — no public IP, no webhook, no hosting platform. It runs on a laptop or a small VPS.

Polkadot app (phone) <-> Statement Store <-> bot-core <-> brain

The "brain" is the AI agent behind the bot: a coding-agent CLI that bot-core drives directly, or an agent framework connected through a small HTTP bridge.

Features

Bring your own agent — replies come from Claude Code, Codex, Kimi Code, or opencode (which reaches Anthropic, OpenAI, Google, local models, and more), or from an agent framework such as Hermes or OpenClaw.
Private by default — conversations are end-to-end encrypted, and a bot answers only the accounts you allowlist unless you explicitly make it public.
A real agent, not a chat wrapper — per-conversation session memory, file attachments in both directions, /project to point a conversation at a repo or git worktree, and in-chat commands like /model and /usage.
One-command deploy — pca deploy ships the bot to any SSH + Docker host, with status, logs, and stop to operate it.
**T3ams s