An AI meeting assistant that knows what you know.
Most AI meeting tools know what was said in the call. Kennan also knows what you’ve already written. Drop in your resume, your strategy doc, your customer brief, your project spec, your latest 1:1 notes. Kennan reads them, indexes them on your Mac, and surfaces the right fragment the moment a conversation needs it.
Generic AI knows generic things.
Plain frontier models can quote the famous interview frameworks back at you. They cannot quote your last quarterly review, the specific deal memo from August, or the line your manager wrote about the launch slip in week three. Kennan’s knowledge base is the difference.
- 01
Drop the file in.
PDF, Markdown, plain text. Resume, deal memo, project doc, customer brief, book. Anything readable.
- 02
It indexes on your Mac.
Kennan builds an embedding of the file on the device. Your documents do not leave the machine to get indexed.
- 03
The conversation pulls the snippet.
When something in the call matches a fragment in your library, the assistant surfaces the right sentence on your HUD with the file name attached. Cited, not invented.
What you can drop in.
Anything readable on a Mac is fair game. The shape of the document does not matter; the assistant works from what is inside it.
-
Resume.
Behavioral interviews quote the project, not the title. Kennan pulls the right project from your resume and surfaces it in STAR shape on the HUD.
-
Strategy doc.
Quarterly plans, OKRs, the deck you wrote for the board. The assistant cites the relevant section when leadership asks where the timeline came from.
-
Customer brief.
Account research, contract history, the deal memo. On a renewal call, Kennan reminds you of last quarter's escalation before you walk into the question.
-
Product or project spec.
The spec your team wrote, the design doc, the API reference. When a stakeholder asks why a constraint exists, Kennan cites the doc that explains it.
-
Meeting notes & 1:1 history.
Last week's review, the running 1:1 doc, the retro notes. The assistant remembers the conversation thread you already had so you do not have to.
-
Reference books & long-form.
A negotiation book, a management playbook, a domain handbook. Drop the PDF in; Kennan can cite a passage word-for-word in the moment.
Where it’s worth it.
Scenarios where having your own context in the room changes the answer.
-
Behavioral interview prep.
Your resume is in the knowledge base. The interviewer asks 'walk me through the API platform project at Acme.' Kennan surfaces the bullets — what you led, the numbers, the outcome — in STAR shape on the HUD.
-
Sales objection handling.
Your customer brief and the deal memo are loaded. The prospect says 'your pricing is too high.' Kennan pulls the comparable-deal data from the brief and works it into the LAER acknowledge + explore line.
-
Performance reviews.
Your running 1:1 doc and the project retro are loaded. Your manager asks where the team slipped on the Q2 timeline. Kennan cites the specific blocker that was logged the week it happened.
-
Cross-functional escalations.
Your product spec and the architecture doc are loaded. A stakeholder asks why a constraint exists. Kennan cites the section of the doc that explains the tradeoff.
Private by architecture.
Your documents are sensitive. Kennan’s knowledge base is built on the assumption that they should never become someone else’s training data.
-
On-device embedding.
The model that turns your documents into searchable vectors runs on your Mac. The raw files and the embeddings live in local storage.
-
Not training data.
We do not collect your documents, your embeddings, or your queries. Kennan has no server in the request path for the knowledge base.
-
Snippet, not file.
If you have pointed Kennan at a frontier model (OpenAI, Anthropic, Gemini), the model sees only the small retrieved snippet that matched the conversation. Never the whole file.
Bring your own knowledge to the call.
Kennan is a real-time AI meeting assistant for Mac, with a personal knowledge base built in. Native Swift, on-device, lifetime license. Join the waitlist to get notified at launch.