a-prs

CliClaw — AI skill for Claude Code

AI community

Universal Telegram AI assistant.

How to install CliClaw

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

What CliClaw does

Universal Telegram AI assistant. Choose backend: Claude Code, Codex CLI, or Qwen Code. Voice, sessions, memory vault, scheduler.

Alternatives in AI

Qwen Audio Agent — A realtime voice runtime that keeps Agents talking, working, and present 2.3k ★
Loushang — AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool 1.3k ★
Registry Broker Skills — AI agent skills for the Universal Registry - search, chat, and register 72,000+ agents across 14+ protocols 410 ★

README

CliClaw

Universal AI assistant in Telegram. Choose your backend: **Claude Code**, **Gemini CLI** (free), or **Codex CLI**.

Features

Multi-backend: Claude Code, Codex CLI, Qwen Code — choose at install
Voice messages: Groq Whisper API (free)
Session management: create, switch, close — inline buttons
Memory vault: auto-saves facts from conversations, injects context from past sessions
Scheduled tasks: natural language cron ("remind me tomorrow at 14:00")
Self-update: /update command from Telegram
Image support: send photos for vision analysis
One-command install on Ubuntu/Debian VPS

Install

curl -fsSL https://raw.githubusercontent.com/a-prs/CliClaw/main/install.sh -o /tmp/install.sh && sudo bash /tmp/install.sh

Or with backend pre-selected:

sudo bash /tmp/install.sh --backend=claude

Backends

Backend	Auth	Cost
Claude Code	API key	Paid (Anthropic API)
Gemini CLI	API key	Free tier (10-50 RPM, no credit card)
Codex CLI	Phone login or API key	Paid (OpenAI API / ChatGPT subscription)

Commands

/menu — control panel
/new — new session
/sessions — session list
/status — system status (backend, memory, voice)
/setup — configure voice, API keys
/update — update bot from GitHub

Memory

The bot automatically:

Saves session logs to workspace/memory/sessions/
Extracts explicit "remember" requests to workspace/memory/facts.md
Searches memory before each prompt and injects relevant context

FTS5 full-text search — finds facts in <1ms.

Server Requirements

1 vCPU, 512MB+ RAM (swap auto-created if <1.5GB)
Ubuntu 22.04+ / Debian 11+
No GPU needed

Architecture

bot/
  main.py          — Telegram bot (aiogram 3.x)
  runner.py        — Generic CLI runner with queue
  backends/        — Claude, Codex, Qwen strategies
  memory/          — Vault (markdown files) + FTS5 search + hooks
  voice.py