Kanyon — a calm desk overlooking a canyon at sunset

Most founders, recruiters, and operators live in their inbox but manage their relationships somewhere else — a spreadsheet, a Notion table, a CRM nobody updates. Kanyon exists because that split never made sense. The inbox already has every person you've talked to, what they said, and when. It just doesn't know what to do with any of it. We built Kanyon to fix that.

The Product: Email That Already Is a CRM

Kanyon is an email client that treats every thread as a deal, every sender as a contact, and every conversation as a point on a timeline. You connect a Gmail or Outlook account, and a few minutes later you're looking at your inbox organized as a pipeline — not a stream.

Instead of a single endless list, you see columns: New, In Process, Interested, Not Interested, Completed, and anything else you want to add. Emails flow into the right column automatically based on content, context, and how you've handled similar threads in the past. When you act on one — reply, schedule, dismiss — it moves. You never have to tell a separate CRM about anything.

Kanyon board view showing email organized as a pipeline

Why We Built It

The problem we kept watching people run into was always the same shape. Someone's doing real relationship work — sales, hiring, fundraising, client delivery — and their source of truth is their inbox. But to keep anyone else informed (or to remember what they themselves said last month), they have to either:

Copy-paste threads into a CRM that immediately goes stale, or
Maintain a spreadsheet that nobody but them can read, or
Give up and trust their memory, which is where most follow-ups quietly die.

Every one of those options loses information. The emails already contain everything a CRM would store, only better — they have timestamps, they have tone, they have the exact words the other person used. Why move them somewhere worse?

Kanyon's bet is that the inbox is the system of record. It always has been. It just needed a better interface on top.

How It Works Under the Hood

Inbox Ingestion Without a Middleman

Kanyon connects directly to your Gmail or Outlook account through the standard sign-in those services already use. No one sits in the middle reading your mail. Permissions are limited to read-only by default, and you can disconnect from the same screen you set it up on.

Once connected, Kanyon pulls your recent inbox history and begins classifying threads in the background. The first full sync typically takes under a minute for a heavy inbox.

Smart Categorization That Learns From You

The categorization engine is the part that makes Kanyon feel alive. When a new email arrives, it gets scored along a few axes: is this a real person or a newsletter, does it reference a previous thread, does it contain an ask, is it waiting on you or the other side? That score plus the sender's history places the email in the right pipeline column.

Every time you move a card, ignore a suggestion, or mark something as "not interested," the engine adjusts. Within a couple of weeks most users report the inbox starts placing about 85–90% of threads correctly without intervention. The remaining 10% is the fun work — the judgment calls — which is exactly the part that should still require a human.

Contacts Inferred From Conversations, Not Imported

We deliberately don't ask you to import Google Contacts. People's contact lists are almost always a mess of old phone numbers, half-filled entries, and strangers. Kanyon instead infers contacts from mailbox interactions: who you email, who replies, how often, how recently. That yields a clean, high-signal contact list that actually reflects the people in your working life right now.

Every contact gets a timeline — every email, every note, every status change — and a Notes tab for the stuff that doesn't belong in an email but you'd lose if you wrote it on a Post-it.

The AI Assistant, Scoped Carefully

Kanyon has a chat assistant that can do things like "draft a polite follow-up to everyone in Interested who I haven't heard from in two weeks" or "find Jane Doe's email at NYU." For the email lookup, we only use verified public sources — if a real page on the public web doesn't literally show the address, the assistant says so and stops. No guessing, no making up contact info. Every result it does return includes the source link so you can check.

That "never bluff" stance is a deliberate design choice. In relationship tools, a wrong answer is worse than no answer.

Design Choices That Took a While to Get Right

The Board View, Not a List

We tried a list with status chips first. It was fine. It was also indistinguishable from every other "smart inbox." The board view is what made the product click for testers — seeing "Interested" and "In Process" as visual buckets, at a glance, triggered the same mental model people use in Trello or a hiring pipeline. Email is already work, it just hadn't been shown that way before.

Never Sync What You Don't Need

Most CRMs brag about how much they import. Kanyon brags about how much it doesn't. We don't pull attachments we don't need to display. We don't copy your mailbox to our servers. We don't ask for write access until a feature literally requires it (drafting, archiving). Every permission you grant should be traceable to a feature you actually use.

What's Next

Right now we're deepening three things: better multi-inbox support for people who run more than one business out of Gmail and Outlook simultaneously, a stronger scheduling integration so a thread can become a calendar hold without leaving Kanyon, and team workspaces for small teams that share a relationship surface.

The north star hasn't changed. Email is where the relationships live. Kanyon just gives them a home they actually belong in.

Try it at kanyon.email.

Building Kanyon: The Email Client That Thinks Like a CRM