Images, web & voice

Beyond chatting and remembering, your agent has optional features you can toggle in the dashboard (Model & features tab). Disabled features cost nothing.

Images and video

With a fal.ai key (added during creation or later in the dashboard), the agent can:

• Generate images from a description (“create a minimalist logo for…”)
• Edit images you send
• Animate images and generate short videos

Results arrive directly in your Telegram chat. Billing goes through fal.ai, per generation — the platform adds no markup.

Generated images and videos remain available for 30 daysand are then deleted automatically so they don’t fill up the agent’s disk. To keep a result forever, just ask: “save it to the library”.

Web browsing

Lets the agent read pages, search the web, and take screenshots of sites. Essential for requests like “look up prices for X” or “summarize this article” (by sending a link).

Audio — two different things

Don’t confuse the two directions of voice interaction:

• Understanding your voice messages (available now): you send a voice message, the agent transcribes it and replies. This uses the transcription service from Step 5 of setup (Groq recommended, with a generous free tier).
• Replying in voice:the agent sends its reply as an audio message when you ask (“reply in audio”). Enable it during agent creation (Step 6, optional) or anytime in the dashboard under Credentials — pick the voice provider (OpenAI, Groq or fal.ai) and enter the API key. If you already use Groq for transcription, the same key works for voice.

Skills

When the agent solves a complex task, it can turn the process into a reusable “skill” — next time, it executes faster and consistently. Manage skills with /skill in chat.