Changelog

We’ve added a number of new models to Chatling, giving you more choice across quality, speed, and cost.

• GPT-6 Astra (25 credits): OpenAI’s most capable model. Strong reasoning, instruction following, and image understanding for complex, multi-step conversations.

• GPT-6 Sol (5 credits): A balanced everyday model. Strong reasoning, image understanding, and tool use at a much lower cost than Astra.

• GPT-6 Luna (1 credit): A fast, low-cost model for high-volume agents. Handles straightforward questions with reasoning, image understanding, and tool use.

• Claude Opus 5.5 (10 credits): Anthropic’s latest Opus model. Strong reasoning and instruction following for detailed, multi-step answers, with image understanding.

• xAI Grok 4.7 (5 credits): A fast reasoning model for detailed questions and longer conversations, with a useful balance of quality and cost.

• GLM 5.3 (3.5 credits): An advanced reasoning model with strong instruction following and context retention. Best for accurate, detailed answers.

• GLM 5.3 Flash (0.5 credits): A fast model with strong reasoning, image understanding, and low latency. Built for high-volume agents.

• DeepSeek V4.1 Flash (0.5 credits): A fast multimodal model with long-context understanding and image support. Suited to detailed inquiries at a low cost.

All models are now available in paid plans. Head to your agent’s AI settings to try them and find the right balance of quality, speed, and cost for your needs.

Your AI agent can now take care of the busywork. With our new visual workflow builder, you design a process once and your agent runs it automatically, every time.

Workflows kick off on their own when something happens in your AI agent, whether that's a new conversation, a contact update, or another event. From there, they can:

• Send emails and add internal notes

• Sync data to HubSpot, Zendesk, Google Sheets, Airtable, and other platforms

• Update conversations and contact details

• Assign conversations to the right teammate or team

• Route tasks based on conditions you set

• Connect to external services with HTTP requests

• And much more

The result is fewer manual tasks, smoother handoffs, and more time for your team to focus on customers.

To create your first workflow, open your AI agent and go to the Workflows page. For more information, refer to the Workflow docs.

Customers no longer need to keep your website open while waiting for a response.

Web widgets can now collect a customer’s email during handoff. If your team replies after the customer leaves, they’ll automatically receive the replies via email.

This feature helps you stay in touch with customers who drop off before your team can respond, so live chat conversations don’t go unanswered.

To configure this feature, head to the Widget Settings → Configure → Email reply notifications.