43 frontier AI models.
One bot. Zero dollars.
Claude, GPT, Gemini, Grok, DeepSeek, Qwen, Kimi and GLM — routed through one OpenAI-compatible gateway and delivered inside a Telegram chat. Ask a question, generate an image, cut a video, or hand a task to an agent that can browse the web.
Models you actually wanted to try
Every model is reached through the same endpoint, so switching providers is a dropdown, not a new account, a new card and a new rate limit.
Four things, one entry point
Chat
Type a question in the bot or the Mini App. Pick a model, get an answer. No registration form, no card, no per-provider account.
Images & video
Text-to-image and text-to-video generation are wired to the same credit balance, with queue status shown right in the chat.
Agent
An autonomous mode with web search and a sandbox: give it a goal, it plans, runs the searches and returns a written result with sources.
Free API keys
Every user can mint an OpenAI-compatible key and point any client at the gateway — the same models, the same limits, no separate billing.
Drop-in OpenAI-compatible endpoint
If your code already speaks the OpenAI format, nothing changes except the base URL and the key.
curl https://api.reformboss.com/v1/chat/completions \
-H "Authorization: Bearer $IISHKO_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"claude-sonnet-5","messages":[{"role":"user","content":"hello"}]}'
Why it exists
Model access is fragmented on purpose: nine providers, nine dashboards, nine cards. иишко collapses that into one Telegram chat run on a shared gateway, with fair-use limits instead of subscriptions. The companion channel @AlStack publishes the free tiers and trial credits that make it possible, verified before every post.
