TeleBotHost

AI for TBL bots

Agents help. You stay in charge.

TeleBotHost works with coding agents and chat assistants that understand TBL. Draft commands, wire keyboards, and manage bots when tools are connected. You review every change before it goes live.

TBL is sandboxed JavaScript for Telegram bots on TeleBotHost. It is not Telegraf, grammY, or a long running Node process. Good AI setups know that from the first reply.

Use the Agent Skill inside Cursor, Claude Code, and similar IDEs. Self host MCP from our GitHub repo when you want the assistant to call the Developer API (public endpoint coming soon). Open the editor extension when you prefer local files. For quick questions in a browser, try TBL Helper AI in Gemini or ChatGPT.

Agent docs map: docs.telebothost.com/for-agents and llms.txt . Same page at /ai .

TBL Helper AI

Chat helpers on Gemini and ChatGPT

TBL Helper AI is available on Google Gemini and ChatGPT. Use it to sketch commands, Answers, keyboards, and Logic. Like every AI surface here, read the output before you apply it in the console, MCP, or editor.

All AI surfaces

Coding agents, API tools, the editor, and browser chat. Pick what matches your workflow.

TBL Agent Skill

Live

Best when an AI coding agent is writing or editing TBL in your project.

An installable Agent Skills pack for coding agents. It teaches hard rules for TBL: top level Logic wrapped in an async IIFE, platform command fields, Api vs Bot, storage scopes, and a one page docs workflow.

  • Works with Cursor, Claude Code, Codex, Copilot, OpenCode, and other Agent Skills clients
  • Stops agents from inventing npm stacks or long running listeners
  • Points agents at docs.telebothost.com/for-agents and llms.txt

Install: npx skills add telebothost/tbl-skill

View tbl-skill on GitHub

TeleBotHost MCP

Live

Best when the assistant should touch live bots, not only draft code in chat.

A Model Context Protocol server for the TeleBotHost Developer API. Compatible assistants can list bots, create or update commands, read logs, manage env vars, and more over HTTP. For now you self host from the open source repo. A public hosted endpoint will be provided soon.

  • Clone and deploy from github.com/telebothost/mcp-server
  • Set TELEBOTHOST_API_KEY (sk_…) and point your AI client at your MCP URL
  • Pass X-Tbh-Api-Key per request when your client supports headers
  • Pair with the Agent Skill so tool calls still produce correct TBL
  • Public shared endpoint coming soon

Self host for now. Public endpoint coming soon.

View mcp-server on GitHub

VS Code / Cursor extension

Live

Best when you want file based editing plus AI in VS Code or Cursor.

Download a bot into a local folder, edit commands/ as files, upload only what changed, and view logs without leaving the editor. On download it adds AGENTS.md, DOCS.md, and CLAUDE.md for agents in that workspace.

  • Marketplace and Open VSX: search TeleBotHost
  • Versions and changelog on the extension page
  • Works well with the Agent Skill in the same editor
Open extension page

TBL Helper AI on Gemini

Live

Best for quick TBL questions when you already live in Gemini.

A Google Gemini Gem tuned for TeleBotHost and TBL. Ask for command fields, Logic snippets, keyboards, and flow ideas in the Gemini chat UI. Like any AI, check the output before you ship it.

  • Open the shared Gem link (Google account required)
  • Ask for name, answer, keyboard, and code as separate fields
  • Apply results in the console, MCP, or editor after you review

TBL Helper AI on Google Gemini

Open Gemini Gem

TBL Helper AI on ChatGPT

Live

Best for quick TBL questions when you already live in ChatGPT.

A custom GPT for ChatGPT that helps with TeleBotHost commands and TBL patterns. Same idea as the Gemini Gem: draft in chat, then apply what looks right.

  • Open the GPT link (ChatGPT access required)
  • Request runnable TBL and clear command fields
  • Paste into the console or push via MCP / extension after you review

TBL Helper AI on ChatGPT

Open ChatGPT GPT

How to combine them

Write TBL correctly

Install the Agent Skill in your coding agent. In chat, use TBL Helper AI on Gemini or ChatGPT for drafts you will review.

Manage bots from AI

Self host the MCP server from the GitHub repo, connect it with your sk_… key, and let the assistant list bots, create commands, and read logs. A public endpoint is coming soon.

Edit as files

Use the VS Code or Cursor extension when you want a local commands/ tree, pending uploads, and logs beside your editor AI.

You approve what ships

None of these tools replace judgment. Read the Logic, check answers and keyboards, then apply changes yourself.

Good to know

What is TBL Helper AI?

TBL Helper AI is our chat facing assistant for TeleBotHost, available as a Gemini Gem and a ChatGPT custom GPT. Use it to draft commands and Logic. Every AI can miss details, so review before you apply changes.

Skill vs MCP vs chat?

The Skill teaches coding agents how to write TBL. MCP lets compatible assistants call the Developer API (self host from the mcp-server repo for now; a public endpoint is coming soon). Gemini and ChatGPT helpers are browser chat for ideas and sample code. Many people use Skill plus MCP together in Cursor.

Will AI invent Telegraf or npm bots?

Without guidance, often yes. That is why the Skill, AGENTS.md from the extension, and TBL Helper AI exist. They steer answers toward sandboxed TBL on TeleBotHost. Still read the output. All models make mistakes.

Where are the docs agents should open?

https://docs.telebothost.com/for-agents/ maps intent to one page. https://docs.telebothost.com/llms.txt lists every path. Agents should open one page per task, not crawl the whole site.

Questions about setup? Contact us or join @update_chat .