Contents
0%The Chat node creates interactive conversational interfaces that allow users to have real-time conversations with your AI-powered applications.
What it does
The Chat node supports both interactive conversations and automated flow runs. In the editor, you can ask follow-up questions, refine requests, and build a reusable sequence of user messages. When the node runs, it replays those user messages in order as a fresh conversation.
Grouping conversations
A Chat node can hold several independent conversations. Use the tabs in its purple header to switch conversations, press + to create a blank conversation, or open a tab's menu to rename, move, or delete it. Long tab lists have horizontal scroll buttons. Use the left and right arrow keys (or Home and End) to focus a tab, then press Enter or Space to open it. Switching tabs preserves each conversation's draft.
- Drag one Chat node onto another, review the destination and connection changes, then confirm to group every conversation from the dragged node into the destination node. Cancel keeps both nodes and their connections.
- Drag a conversation tab onto another Chat node to move only that conversation.
- Drag a tab onto blank canvas, or choose Move to new Chat node, to extract it into a new node without connections.
Grouping never concatenates messages. The destination keeps its label, position, size, model, tools, and existing connections. The dragged source node and its connections are removed. An extracted node copies the source model, tools, and dimensions but starts without connections.
Conversation organization is unavailable while the affected Chat node is generating or replaying.
Deleting any conversation requires confirmation. Deleting the final conversation also removes its Chat node and all of that node's connections.
Content
- Text from connected parent nodes (e.g., Text, Web, File)
Output
- During an automated run, the node outputs a complete transcript containing each saved user message and its newly generated assistant response.
- The run appears live inside the Chat node: each saved user message is added when its turn starts, followed by the streamed assistant response.
- Connected parent-node outputs are included as context for the replay.
- Linked reusable groups are resolved from their source group at run time, so reruns use the latest source data.
- Responses from the authored conversation are not reused. Each run generates fresh responses and carries each new response into the next turn.
Settings
- Model: All conversations in a Chat node share its current model. During grouping, the destination node's model wins.
- Tools: All conversations in a Chat node share its selected tools. During grouping, the destination node's tools win.
Automating a conversation
- Add a Chat node, select the conversation tab you want to automate, and write the user messages you want to replay.
- Connect any parent nodes whose output should be available as context.
- Use Run This to replay only the Chat node, or Run All to run the connected flow.
Only the displayed conversation is replayed. A blank displayed conversation succeeds with an empty replay and empty output. If any turn fails, the Chat node stops immediately and downstream nodes do not run.
Credits cost
Each saved user message in the displayed conversation is one generated turn. A replay therefore costs the selected model's credit cost multiplied by that conversation's saved user-message count; messages in the other tabs are not replayed or charged. Completed turns remain charged if a later turn fails.
| Name | Credits |
|---|---|
| GPT-4.1 nano | 1 credits |
| GPT-4.1 mini | 5 credits |
| GPT-4.1 | 10 credits |
| GPT-4o | 20 credits |
| o3 | 25 credits |
| o3 pro | 300 credits |
| o4 mini | 15 credits |
| GPT-5 nano | 1 credits |
| GPT-5 mini | 5 credits |
| GPT-5 | 10 credits |
| GPT-5 Chat | 10 credits |
| GPT-5 Pro | 350 credits |
| GPT-5.1 | 10 credits |
| GPT-5.2 | 14 credits |
| GPT-5.4 | 18 credits |
| GPT-5.5 | 36 credits |
| GPT-5.6 Luna | 8 credits |
| GPT-5.6 Terra | 18 credits |
| GPT-5.6 Sol | 36 credits |
| Gemini 2.5 Flash | 10 credits |
| Gemini 2.5 Pro | 150 credits |
| Gemini 3 Pro | 200 credits |
| Gemini 3 Flash | 10 credits |
| Gemini 3.1 Flash Lite | 5 credits |
| Gemini 3.1 Pro | 220 credits |
| Claude Haiku 4.5 | 20 credits |
| Claude Sonnet 4 | 50 credits |
| Claude Sonnet 4.5 | 55 credits |
| Claude Sonnet 4.6 | 55 credits |
| Claude Opus 4 | 300 credits |
| Claude Opus 4.5 | 80 credits |
| Claude Opus 4.6 | 80 credits |
| Claude Opus 4.7 | 160 credits |
| Claude Opus 4.8 | 160 credits |
| Kimi K2 Instruct | 10 credits |
| Grok 4 Fast Reasoning | 6 credits |
| Grok 4 Fast Non-Reasoning | 4 credits |
Use cases
Customer Support
- Help Desk Bots: Automated customer service with escalation
- FAQ Systems: Interactive Q&A with contextual follow-ups
- Troubleshooting Guides: Step-by-step problem solving
Interactive Applications
- AI Assistants: Personal or business productivity helpers
- Learning Companions: Educational tutoring and guidance
- Creative Collaborators: Brainstorming and idea development
Business Tools
- Lead Qualification: Interactive sales and marketing bots
- Consultation Tools: Guided decision-making processes
- Data Collection: Conversational forms and surveys
Specialized Services
- Content Generators: Interactive writing and creation tools
- Analysis Tools: Conversational data interpretation
- Planning Assistants: Interactive strategy and planning
Continue Reading
Discover more insights and updates from our nodes
Upload and process various file formats in your workflow for document analysis and content extraction
Generate high-quality images using AI models based on text descriptions and prompts
Upload, store, and manage images in your workflow for AI analysis and processing