Index / thepopebot · updated Oct 2, 2026
thepopebot
stephengpope/thepopebot · healthy · rank 22 of 43 by stars
A self-hosted personal agent platform that unifies chat (Telegram/web), any LLM, and any coding agent into one app with live browser terminals and background Docker jobs. It leans hard into 'real work, not just chat' — agents open PRs, auto-merge, and DM you when done.
Facts
Repository
Runtime
Posture
Community
Security breakdown
Composite 72 / 100 · how these are scored
higher is safer
higher is safer
higher is safer
higher is safer
higher is riskier
Model access
Read from the repository, not written by a model · 63 files examined
5 providers · runs locally · gateway support
Pinned models
None pinned in the repository.
Evidence
Decision
Why choose thepopebot over OpenClaw?
Why choose this
- Unified chat + live terminal + background job model in one app
- Works with any coding agent (Claude Code, Codex, Gemini, etc.) out of the box
- Docker-isolated background jobs with PR auto-merge workflow
Tradeoffs
- Smaller ecosystem and fewer integrations than OpenClaw
- Only Telegram supported today; Slack/Discord not shipped
- Single-maintainer project with bus-factor risk
Best fit
- Solo devs wanting a 24/7 autonomous coding agent
- Users who want Telegram-driven agent workflows
- Self-hosters who want PR-open/auto-merge automation
Avoid if
- You need multi-user team collaboration features
- You want a managed cloud service with zero ops
- You need Slack/Discord integrations today (still coming soon)
Evidence is solid from README, commits, and release history showing active development, but there is zero Reddit discussion and web sentiment is indirect (mostly about OpenClaw generally), so community and security assessments carry moderate uncertainty.
AI layer reviewed Sep 21, 2026 · how this is written
Star activity
1,862 stars today
Overview
Overview
thepopebot is a self-hosted personal AI agent platform that combines a chat interface, a live coding environment, and a background job runner into a single Node.js application. Its core architecture is described as 'three doors, one brain': browser chat, Telegram (with per-user verification), and code workspaces all feed into a central event handler that picks the coding agent, manages sessions, and orchestrates Docker containers. It is LLM-agnostic (Anthropic, OpenAI, Google, DeepSeek, OpenRouter, and any OpenAI-compatible endpoint) and agent-agnostic (Claude Code, Codex, Gemini, OpenCode, and others).
The standout differentiator is the split between live chat (agent runs in-process or in a headless container, streaming to your screen) and agent jobs (background workers that spin up fresh Docker containers, open PRs, auto-merge, and DM you on completion). This makes it closer to an autonomous junior developer than a chatbot. Recent commits show active, careful maintenance — fixing stale config caches that served outdated bot tokens, adding Telegram webhook URL overrides for proxied deployments, and tightening the admin UI.
Compared to OpenClaw, thepopebot is narrower but more opinionated: it focuses on the coding-agent-plus-chat loop rather than broad multi-channel assistant features. Everything runs on your hardware with your tokens, giving it a strong privacy posture, though unrestricted shell access via live terminals and Docker jobs means operational risk is non-trivial. Slack and Discord integrations are promised but not yet shipped.