Context-aware replies
Uses the selected message, up to 15 recent turns, and available memory for the active contact. History completeness depends on the site.
AI assistant for web chats
No time to type? Chatwing AI helps you understand the message and draft a fast, natural, context-aware reply directly on the conversation page.
A test build for Google Chrome. Extract the archive and load the folder from chrome://extensions with Developer mode enabled. Chrome Web Store publication is still being prepared.
Test version 0.1.90. Cloud AI needs only a license from the Telegram bot; no API key is required.
For the test build: install it, connect access, consent to conversation processing, and test your selected model.
Open @chatwing_auth_bot for trial access or a subscription. The bot provides a license. No separate API key is needed.
In Settings, paste the License key and select Check subscription. Alternatively, create a Telegram code, send it to the bot, then select Check link.
Consent to conversation processing. Choose OpenRouter and run Test AI connection. Model selection is automatic.
Select a message and open the panel. Reply or Generate drafts a context-aware response; TL translates, and Improve revises the text into English. Review before sending. If the site blocks insertion, copy the draft.
Chatwing AI drafts replies, translates text, and keeps separate local memory for conversations. Messages are sent only after your action on the site.
Uses the selected message, up to 15 recent turns, and available memory for the active contact. History completeness depends on the site.
TL translates selected text. Improve corrects and expands it into English; its default prompt is editable in Settings.
Cloud AI runs through the Chatwing server and OpenRouter within subscription limits. Personal API keys are encrypted on the server and are not issued to clients.
Settings offer free text models, budget paid models and Claude Sonnet.
Subscription and trial access are issued through the Telegram bot and checked by the license server.
History, summaries, and spending logs stay in Chrome. Optional semantic memory sends message text to OpenRouter Embeddings when enabled.