Pepebot is an ultra-lightweight framework to run AI assistants across 200+ LLMs, multiple messaging platforms, and full system automation — all from a single binary.
$ pepebot gateway
[INFO] Starting Pepebot v0.x …
[INFO] Loaded LLM: claude-sonnet-4-6
[INFO] Telegram channel active
[INFO] HTTP gateway listening on :18790
$
200+
LLM Models
5+
Messaging Channels
8
Platforms Supported
MIT
Open Source
Built for developers who want a capable AI assistant without the bloat.
Switch between 200+ models via MAIA Router. Supports Anthropic, OpenAI, Groq, Gemini, Zhipu, and more.
Execute shell commands, manage files, and trigger web searches — all through natural language instructions.
Automate Android devices with ADB. Record activities and replay them as reusable workflows.
Multi-step workflows with variable interpolation. Compose complex tasks from simple building blocks.
Receive and transcribe voice/audio messages automatically across supported chat channels.
Built-in HTTP server on port 18790 for external integrations and the web dashboard.
Connect Pepebot to your favorite messaging platform.
LLM Providers
Multiple installation methods to fit your workflow.
One-liner
Auto-detects your OS and architecture.
Homebrew
brew tap pepebot-space/pepebot
brew install pepebot
macOS and Linux via Homebrew.
Docker
Also supports Podman.
Pepebot is free, open source, and runs entirely on your own infrastructure. No subscriptions, no telemetry.
View on GitHub