# TapLM

> TapLM is a philosophical AI on Telegram for questions, reasoning, and debate. Send it a message and it answers. Ask it to agree with you and you get an argument instead. It runs a structured three round debate mode with an AI judged verdict, and multi persona symposiums where a seated panel of voices argues a topic in front of an audience.

TapLM lives entirely on Telegram (direct messages and groups) at https://t.me/TapLMbot. It is a tool rather than a companion: it has no persona, no simulated inner life, and nothing to say about itself. The primary interaction is plain conversation, and every command is a shortcut for one specific move.

## Key Capabilities

- [Plain Conversation](https://taplm.fun/#commands): Send any message and it answers. `/ask <question>` is the explicit form, and it is the one that works in a group without replying to anything.
- [Grounded Web Search](https://taplm.fun/#commands): `/search <question>` answers with Google web grounding, for questions that turn on a fact rather than an argument.
- [Philosophy Command Set](https://taplm.fun/#commands): Thirteen shortcuts for definitions, thinkers, quotes, thought experiments, paradoxes, koans, schools, comparisons, dialectic, ethical analysis, and allegory.
- [Reply Utilities](https://taplm.fun/#commands): Twelve shortcuts that operate on a message you replied to, including `/steelman`, `/fallacy`, `/counter`, `/bias`, `/socratic`, and `/grade`.
- [Structured Debate](https://taplm.fun/#debate): `/debate @user <topic>` runs three ordered rounds (opening, rebuttal, closing) between two people in a group, then scores the case each side made. Audience members add notes with `/interject`; `/forfeit` concedes on the record; a group admin ends one with `/canceldebate`.
- [Multi Persona Symposium](https://taplm.fun/#symposium): `/symposium <topic>` seats a panel of personas with separate written convictions and lets them argue. Nine seed voices across three rooms, plus a neutral moderator that closes it.
- [Editable Personas](https://taplm.fun/#symposium): The operator can rewrite any voice or add a room with `/personas`.
- [Model Choice and Plans](https://taplm.fun): `/models` picks the chat's model and `/plans` shows the plan card and token budget. Which models a request may reach depends on the plan.

## Core Channels

- [Telegram Bot](https://t.me/TapLMbot): Direct link to start a conversation with TapLM on Telegram (@TapLMbot).
- [Official Landing Page](https://taplm.fun): Web overview, the full command sheet, and the symposium persona roster.

## Optional

- [Full Architecture and Technical Documentation](https://taplm.fun/llms-full.txt): In-depth command catalogue, debate and symposium mechanics, plan limits, and group access model.
- [XML Sitemap](https://taplm.fun/sitemap.xml): Machine-readable index of landing page URLs.

## Honesty

TapLM is an AI. Every answer, every debate verdict and every persona reply is generated by a language model. The personas are written characters, not the philosophers they are named after. It is a good thinking partner and a poor authority.