Yandex AI Studio

You can build and run Telegram bots with pyTelegramBotAPI, deploy Python projects to Yandex Cloud VMs over SSH, publish static sites on SourceCraft Sites, generate PWA files (manifest, service worker, mobile meta tags), and build apps with the OpenAI-compatible Responses API including RAG, streaming, and image generation.

Python developers and engineers building bots, web applications, or AI-powered services on Yandex infrastructure. They suit anyone needing Telegram bot scaffolding, remote VM deployment, static site publishing, PWA conversion, or integration with Yandex AI Studio models, SpeechKit, Vision OCR, and vector stores.

The vm-deploy skill connects to a prepared Yandex Cloud VM via SSH, sets SSH key permissions, installs uv, copies the project to ~/projects/, creates the environment, and generates start, stop, and status scripts. The telegram-bot skill can also trigger remote deployment on request.

It covers client setup for the OpenAI-compatible Responses API, model selection and context limits, text and multimodal responses, streaming, Pydantic structured output, function/web/file search, RAG with vector stores, MCP servers, Code Interpreter containers, image generation, Vision OCR, SpeechKit, and OpenAI Agents SDK integration.

You need Python with uv for local environments, a Telegram bot token for the bot skill, SSH access to a prepared Yandex Cloud VM for deployment, and the SourceCraft src CLI for publishing static sites. The ai-studio skill requires Yandex AI Studio credentials for the Responses API.