Your main way to work with Mimir. Ask questions, generate artifacts, and build Mimir's understanding through conversation.
Mimir's chat is not a generic AI chatbot. Every conversation is grounded in your project's actual data — the sources you've uploaded, the themes that emerged from synthesis, the recommendations Mimir generated, and the business context you've shared. When you ask a question, the AI draws on all of this to give you an answer rooted in evidence, not guesswork.
Think of it as a strategic advisor who has read every customer interview, memorized every data point, and understands what your company is trying to achieve. It can synthesize across sources, connect dots you might have missed, and challenge your assumptions with evidence from your own data. Every conversation also strengthens Mimir's understanding — context you share is automatically extracted and added to your project's knowledge base.
The best way to use chat is to treat it like a conversation with a sharp product advisor. Ask the kinds of questions you'd ask a trusted colleague who has deep context on your product.
Surface patterns
“What are users most frustrated about?” — “Which themes appear in more than half of my sources?” — “Summarize the 3 most important themes.”
Make decisions
“Should I prioritize onboarding or retention?” — “What's the strongest evidence for building a dashboard feature?” — “Compare the top two recommendations.”
Explore ideas
“What would a minimal version of recommendation #3 look like?” — “How would enterprise users react to this change?” — “What are we missing?”
Get specific
“Pull every quote related to pricing from my sources.” — “Which sources mention competitor X?” — “What did the user in the March 12 interview say about mobile?”
Each project in Mimir has its own chat history. When you switch projects in the sidebar, the chat panel switches context too — you'll see the conversation history for that project, and the AI's responses will be grounded in that project's sources and insights.
This means you can have a deep conversation about retention in one project, switch to your onboarding research project, and pick up a completely separate thread without any context bleeding between them.
You can attach files directly in chat using the paperclip icon or by dragging and dropping onto the chat panel. Attached files are processed as sources and automatically incorporated into your project, so they'll influence future analysis runs.
Supported formats
Text files (.txt, .csv, .pdf) and images (.png, .jpg, .webp). Images are processed with Claude vision to extract text from screenshots, dashboards, or whiteboard photos.
When to use it
Attaching files in chat is the fastest way to add a quick source without navigating to the Sources page. Drop in a screenshot of a support ticket, a CSV export from your analytics tool, or a transcript from a customer call — Mimir handles it inline.
Paste a URL into chat and Mimir reads the page content automatically. You don't need to include https:// — bare domains work too. Mimir detects the URL, fetches the page, and incorporates its content into the conversation.
Great for
Sharing competitor pages, blog posts, product announcements, industry articles, or documentation. Instead of summarizing a link yourself, just paste it and ask Mimir to analyze it in the context of your project.
Examples
“What does competitor.com/pricing tell us about their positioning?” — “Read this article and tell me what's relevant to our roadmap: techblog.com/future-of-pm”
Type / in the chat input to open the command menu. Slash commands are shortcuts for common actions you can trigger without leaving the conversation. Most of the time natural language works just as well, but commands are handy when you know exactly what you want.
Artifact commands (/email, /brief, /prd) open a dedicated artifact tab in the workspace panel where the generated document streams in alongside your conversation. You can also request artifacts through natural language — just ask Mimir to write an email, brief, or PRD and it will detect your intent.
Every conversation you have with Mimir makes it smarter about your product. When you share business details — your target market, a competitive threat, a metric you care about, a constraint you're working under — Mimir automatically extracts that information and adds it to your project's knowledge base.
This happens in the background with no extra action on your part. Over time, the more you chat, the deeper Mimir's understanding becomes, and the better its analysis and recommendations get. See how Mimir learns for the full picture, or review what Mimir has learned in business context settings.
Toggle the chat panel
Use Cmd + . on Mac or Ctrl + . on Windows/Linux to quickly show or hide the chat panel without reaching for the mouse.
Validate before you analyze
Use chat to test assumptions and explore hypotheses before running a full analysis. It's faster to ask “Do any of my sources mention churn?” than to re-run the entire pipeline.
Quick source additions
Instead of navigating to the Sources page, attach files directly in chat. Mimir processes them as sources and you can immediately ask questions about the content.
Use it as a sounding board
Before committing to a product direction, describe your plan in chat and ask Mimir to poke holes in it. The AI will reference your actual user data to support or challenge your thinking.
Share context naturally in conversation
The more Mimir knows about your business, the more useful the conversation. Even a few sentences about your product, users, or constraints will noticeably improve response quality — Mimir picks up on context automatically as you chat.