Open-source Grok Bot alternative. Pots is an AI coworker that reads every channel, decides what matters, and acts on its own computer. Self-hosted, any model.
翻译准备中
正在加载完整 README。
查看完整 README ↗The open-source Grok Bot alternative: an AI coworker that acts on its own.
It reads every channel, email and event you give it, decides what matters, does the work on its own computer, and speaks up when it should.
Demo · Compare · Quick start · How it decides · How it works
⭐ If Pots is useful to you, a star helps other people find it.
[!TIP] Setting this up with an AI agent, such as Claude Code, Codex or Cursor? Have it read docs/for-agents.md first. That page explains Pots to your agent and how to set it up. In short: the agent installs it and makes the bot, you sign in to your model (and a judge such as Jev, for group chats) with
pots login, and then the bot connects Slack and your other apps itself.
Grok Bot, dots and Claude Tag showed what an AI teammate can be. pi-durable-bot gives you the same kind of coworker as open source: it runs on your machine, with the model you choose, in the places your team already talks.
Grok Bot, OpenAI dots, Claude Duet, Claude Tag and Meta Muse are closed products. They run on someone else's computer, with someone else's model, inside the apps they chose. pi-durable-bot is MIT-licensed and runs where you put it.
| pi-durable-bot | Grok Bot (xAI) | dots (OpenAI) | Claude Duet (Anthropic) | Claude Tag (Anthropic) | Muse (Meta) | |
|---|---|---|---|---|---|---|
| Source code | Open, MIT | Closed | Closed | Closed | Closed | Closed |
| Where it runs | Your server, your VM, your office box | xAI's cloud | OpenAI's cloud | Anthropic's cloud | Anthropic's cloud | Meta's cloud |
| Model | Any: your subscription login, 40+ providers, or a local model | Grok | OpenAI's | Claude | Claude | Meta's |
| Where you talk to it | Slack, Discord, Telegram, and any channel with a Chat SDK adapter | Mac and iOS apps; Slack for team bots | ChatGPT, Slack, Teams | Its own app | Slack | Web, mobile app, WhatsApp |
| Your data | Stays on your machine, in each bot's folder | Sent to xAI | Sent to OpenAI | Sent to Anthropic | Sent to Anthropic | Sent to Meta |
| Custom work | Anything. Change the code, add tools, connect your own systems | What the vendor allows | What the vendor allows | What the vendor allows | What the vendor allows | What the vendor allows |
The other products as their makers described them in public, November 2026; they may have changed since. Grok Bot, dots, Claude Duet, Claude Tag and Muse are products of xAI, OpenAI, Anthropic and Meta. pi-durable-bot is not affiliated with any of them.
Any model. Log in once with a subscription you already pay for, use an API key from any of the 40+ providers pi supports, point it at any OpenAI-compatible endpoint (vLLM, SGLang, a proxy), or run a local model with Ollama, LM Studio or llama.cpp. The bot uses pi's own model layer, so every model pi can use, the bot can use. With a local model for the agent and the judge, nothing leaves your machine. You are not locked to one model vendor.
Any channel. The core speaks in five plain words: channel, place, thread, message, person. Platform details stay inside one adapter per platform. Slack, Discord and Telegram work today; Teams or email is one more adapter, not a rewrite. One bot can be in many channels and stays the same person everywhere.
Your business. Give it your own tools, data and workflows. It can run several bots on one machine, each with its own name, persona, responsibilities and memory. A bot can even edit its own definition, or create a new bot.
Most AI bots are a command line with a chat bubble. You @ them, they answer, they forget. A real coworker works differently:
| A chat bot | pi-durable-bot | |
|---|---|---|
| What it sees | The message that mentions it | Everything in the places it was invited to: chats, DMs, emails, events, pages it watches |
| When it acts | When you @ it | When it decides it should. It can also answer when you @ it |
| What it remembers | This thread, maybe | One long memory across every channel, kept on disk |
| How it works | Calls the tools someone wired in | Writes and runs its own code on its own computer |
| What happens later | Nothing | It sets its own reminders and comes back |
| After a crash | Starts over, or sends twice | Picks up where it stopped. Never sends the same message twice (on Telegram it can, see why) |
Every message in every place you invite it to lands in one inbox: channel messages that never mention it, edits, deletions, reactions, people joining, and its own messages. It keeps all of it in a local SQLite log and reads that log with SQL whenever it needs context. Nothing is thrown away when a conversation gets long.
It is woken in three ways, the same three ways a person notices work:
以上为 README 的开头部分。其余内容请阅读原文。 查看完整 README ↗