AI Assistant
Using the AI Assistant
The AI Assistant is a built-in conversational tool that lets you ask natural-language questions about your RenderStack data — templates, renders, assets, API keys, activity logs, settings, and platfor
How to Access#
There are three ways to open the AI Assistant:
- Sidebar: Click AI Chat in the left sidebar navigation
- Floating Widget: Click the bot icon in the bottom-right corner (available on all pages except the template editor)
- Keyboard Shortcut: Press the
akey from any page (only works when you're not typing in an input field)
Floating Widget#
The floating chat widget appears as a circular button in the bottom-right corner of your screen. Click it to:
- Quick ask: Type a question and send it — a new session is created and you're taken to the full chat page with the AI's response
- View recent chats: See your 3 most recent conversations and click to continue any of them
- New Chat: Start a fresh conversation
- Open Full Chat: Jump to the dedicated AI Chat page
The widget shows a dot indicator when you have existing chat sessions.
Full Chat Page#
The dedicated chat page at /ai-chat provides the full conversational experience:
Session Sidebar#
- Lists all your conversations ordered by most recent
- Click any session to load its messages
- Delete sessions you no longer need (with confirmation)
- Create new sessions with the "New Chat" button
Message Area#
- User messages appear on the right with a blue background
- AI responses appear on the left with markdown formatting (bold, bullets, headers, code blocks)
- Source citations appear as clickable badges below AI responses
- Processing time is shown for each response
Source Citations#
When the AI references your data, it includes clickable source badges showing:
- The type of entity (template, render, asset, etc.) with an icon
- The entity name
- Click any badge to navigate directly to that item in the app
What You Can Ask#
The AI Assistant can search across all your RenderStack data:
- Templates: "What templates do I have?" / "Show me templates with the tag 'marketing'"
- Renders: "What's the status of my recent renders?" / "How many renders have failed?"
- Assets: "List my uploaded fonts" / "Do I have any SVG assets?"
- API Keys: "Which API keys are active?" / "Show me my GET keys"
- Activity Log: "What happened in the last hour?" / "Show recent errors"
- Settings: "What's my AWS S3 configuration?" / "List my allowed hosts"
- Help: "How do I create a template?" / "Explain dynamic image URLs"
- Responsive Positioning: "How does the percentage reference dimension work?" / "Why does my element move when I change the aspect ratio?"
Responsive Percentage Reference Dimension#
The AI Assistant understands the responsive percentage reference dimension feature. You can ask about:
- Reference dimension basics: "What is the W/H reference toggle?" — Each percentage-based property (X, Y, Width, Height) and each pin offset can resolve its percentage against either the artboard width or height. The default is width-based.
- Troubleshooting layout shifts: "Why does my element move when I change aspect ratio?" — This usually happens because a percentage offset is resolving against a dimension that changes when the aspect ratio changes. Switching the reference dimension (W or H) for the affected property can fix this.
- Uniform spacing: "How do I keep equal spacing from all edges?" — Use width-based reference (the default) for both X and Y offsets. Setting both to 10% means both resolve against the artboard width, producing equal pixel distances from the left and top edges regardless of aspect ratio.
For full details, see the Layers documentation (Transform Units section) and the Template Editor documentation (Transform and Pinning sections).
Tips for Better Results#
- Be specific: Mention template names, asset types, or specific features
- Ask follow-up questions: The AI remembers your conversation context (up to 10 previous messages)
- Use natural language: Ask questions as you would to a colleague
- Reference platform concepts: Use terms like "renders", "layers", "API keys" for more targeted results
- Multi-part questions: You can ask multiple questions in one message — the AI will address each part
AI Access & Availability#
The AI Assistant's availability depends on your organization's plan and add-on status.
Access by Plan#
| Plan Level | AI Availability |
|---|---|
| Free / Trial | Not available |
| Pro | Available as a paid add-on |
| Business | Included in plan |
| Enterprise | Included in plan with higher limits |
Add-On Purchase#
If your plan supports the AI Assistant as an add-on:
- Organization owners can purchase the add-on from the AI Chat page or the Billing page by clicking Add AI Assistant.
- Non-owners will see a message asking them to contact their organization owner.
- Once purchased, the AI Assistant is immediately available.
- The add-on can be removed from the Billing page. It remains active until the end of the current billing period.
Usage Limits#
The AI Assistant has a monthly limit on LLM calls (messages that trigger an AI response):
- Your usage counter is displayed in the chat interface.
- When you reach the usage cap, the chat input is disabled with a "Usage limit reached" message.
- Your existing chat history remains viewable in read-only mode.
- Usage resets at the beginning of each billing period.
Disabled States#
If the AI Assistant is not available, you will see one of these messages:
- "AI not available on your plan": Your plan tier does not support AI. The sidebar icon appears grayed out.
- "Available as add-on": AI can be purchased. Organization owners see a purchase button.
- "Usage limit reached": You've used all your LLM calls for this billing period. Chat history is read-only.
- "AI Assistant is temporarily unavailable": An administrator has disabled AI access for your organization.
Session Management#
- Sessions are automatically titled based on your first message
- All conversations are private and scoped to your account
- Delete sessions anytime — this permanently removes all messages in that session
- Sessions are listed by most recent activity