v1.6 — OpenRouter, Claude & GPT-5.5

Added

• OpenRouter provider support: models can declare provider: openrouter
  in models.yml and are routed through an OpenAI-compatible OpenRouter
  client. New openrouter_api_key / openrouter_api_base config options.
• Anthropic Claude models via OpenRouter: Claude Opus 4.8, Claude Sonnet
  and Claude Haiku.
• OpenAI gpt-5.5 (via OpenRouter).

Changed

• Chat models are now dispatched by their type in models.yml instead of
  hardcoded model-name lists, so adding a model is a config-only change.
• Token counting falls back to the o200k_base encoding for models unknown
  to tiktoken (e.g. Claude).
• The /settings model picker lays buttons out in rows of two to stay within
  Telegram's per-row inline-button limit.
• Image understanding is now driven by a vision: true flag in models.yml,
  so any vision-capable model (GPT-4o, GPT-4o mini, GPT-5.5, Claude) can read
  images — no longer limited to GPT-4o / GPT-4 Vision.

Added GPT Vision Support to Telegram

Changes include:

• Added new model.
• Integrated GPT Vision API calls within the bot's message handling logic.

Usage:

• Send a plain and compressed photo to the chat, don't forget to put a proper caption to it.
• Wait for the response.
• Profit!

What's Changed

• Fix HTTPX problem by @karfly in #314
• Added GPT-3.5 16K chat functionality & Price update by @maccagnit in #306
• fixed grammar and spelling by @Light2Dark in #301
• fix argument of type 'float' is not iterable by @TanNhatCMS in #211
• Allow channels by @Fl0p in #312
• create .dockerignore by @realies in #344
• Make the DALLe image size configurable by @alfaridi in #365
• Store Voice temp files in memory by @ohld in #386
• Vision by @OlegGoless in #442

New Contributors

• @maccagnit made their first contribution in #306
• @Light2Dark made their first contribution in #301
• @TanNhatCMS made their first contribution in #211
• @Fl0p made their first contribution in #312
• @realies made their first contribution in #344
• @alfaridi made their first contribution in #365
• @ohld made their first contribution in #386
• @OlegGoless made their first contribution in #442

Full Changelog: v1.4...v1.5

👩‍🎨 DALLE 2, Group Chats and new 10 new chat modes!

3 big so awaited updates are here. Update your code (git pull) and have fun with new features!

1. DALLE 2:

   • Image generation is implemented as a separate chat mode 👩‍🎨 Artist. You can select it with /mode command
   • You can edit return_n_generated_images parameter in config.yml to adjust number of generated images for each text request
   • Money spent on image generation is carefully calculated (/balance command)

2. Group Chat support:

   • There was an annoying problem – bot answered to all messages when it was added to group chat. Now it's fixed and bot answers only when it's @ tagged or when user replies to bot's message
   • Send /help_group_chat command to bot to get instructions on how to add bot to group chat. Video instruction is also available there

3. New 10 chat modes:

   • Chat mode menu is updated. Fancy pagination is added
   • New chat modes:
     • 🇬🇧 English Tutor
     • 💡 Startup Idea Generator
     • 🧠 Psychologist
     • 🚀 Elon Musk
     • 🌟 Motivator
     • 💰 Money Maker
     • 📊 SQL Assistant
     • 🧳 Travel Guide
     • 🥒 Rick Sanchez (Rick and Morty)
     • 🧮 Accountant

Love. Fun. AI!

🪄 Real-time Message Streaming

I'm excited to announce new feature: 🪄Real-time Message Streaming
Now you don't have to wait until full answer is generated by ChatGPT – messages are streamed word-by-word in real-time.

I hope it will dramatically enhance your messaging experience, because it feels like magic! Just have a look:

🎤 Voice message recognition

Now you can send a voice message and ChatGPT will answer you!

It supports 99 languages and works even better than Telegram voice-to-text premium feature.

Powered by OpenAI Whisper API.

ChatGPT support

Recently OpenAI released their ChatGPT API. It's 10x cheaper and has native ability to maintain the context of the dialog.

Starting with this version ChatGPT API will be the default option for bot. If you want to use older text-davinci-003 API, you can set use_chatgpt_api: false in config/config.yml.

Let's go! 🍿