Skip to main content

What's new

Curated highlights of what we've shipped recently. We move fast and in the open — these are the changes worth knowing about, not every commit.

June 2026

  • Upload ANY image (plain JPGs included) for the launcher button — an in-app drag + zoom crop step previews it masked in the currently selected button shape, then the upload pipeline auto-optimizes (downscale + WebP re-encode; logo uploads also get auto-trim + alpha analysis).

  • Image-only launcher mode: the custom image's alpha silhouette IS the button — no background, border, or shape chrome; glow follows the alpha outline via drop-shadow. A failed image load falls back to the chromed launcher.

  • The entire visitor journey — greeting, AI chat, contact/capture forms, WhatsApp message prefill, confirmations, offline flow — renders in the visitor language (en / es / pt-BR / ar / id / he) with a runtime in-widget language switcher and RTL support.

  • Owners translate their own custom widget copy (welcome messages, first-message template, quick replies) into all enabled languages with one click — AI drafts are shown in an editable review grid and nothing is saved until the owner reviews and saves. Available on the FREE plan.

  • Every scored lead arrives with its reasons in plain words — e.g. "Reached out before · Clicked WhatsApp · Was on your pricing page" — in the dashboard lead card/panel AND in the Slack / Telegram / email / forwarder pings.

  • Repeat visits paint the launcher instantly from a safe local cache of the last good config while the network revalidates in the background; a disabled or schedule-hidden widget can never flash from cache, and open/after-hours status is recomputed at paint time.

  • Agents connect Telegram in one tap — a t.me deep link opens the bot, they hit Start, and the dashboard detects the connection automatically. No bot codes, no chat IDs to find or paste.

  • Business-hours simulator in the widget preview: "Preview at…" chips (Now / Tonight 9pm / Sunday / next configured holiday) feed a simulated clock through the SAME status functions the deployed widget runs, so owners see the open chat vs after-hours form — and the next-open line — exactly as a visitor would at that time.

  • Owners can test their AI in the dashboard preview; when it can't answer, Toran flags the gap and offers an inline input that saves the answer straight into the FAQ knowledge base in one click.

Want something specific? Tell us what would make Toran a no-brainer for your business — we read every message.