Why Limn?

The case for keeping the conversation next to the text

Every team already has a place to write. What they're missing is a place to think out loud about what they wrote — without opening a pull request, without scheduling a meeting, and without digging through a Slack thread three weeks later to remember the decision.

Limn puts that conversation inside your editor, anchored to the exact line it belongs to. Your Claude Code agent reads every thread. Your teammates reply from any browser. The history stays in the repo, forever.

The four objections we hear most

1. "We already use GitHub PR comments."

PR comments are gated behind a diff. They live on a specific commit, not on the file. Close the PR and the conversation is archived, severed from the living document. Limn threads are attached to the file path and anchor text — they survive refactors, merges, and rebases. You can start a thread on a design doc before a single line of code is written, and it's still there when the code ships.

Limn is for ongoing discussion. PRs are for point-in-time review.

2. "We use Notion for this."

Notion is great for pages that belong in Notion. But your architecture decisions, onboarding guides, and runbooks already live in Markdown files next to the code. Moving to Notion means a second place to maintain, a context switch every time someone wants to comment, and a growing gap between what the doc says and what the code does. Limn keeps the discussion where the document lives.

Limn is for the files already in your repo, not a second wiki.

3. "Our company uses Confluence."

Confluence is the right tool for policy documents that require approval workflows and enterprise search. It is a poor tool for developer-facing docs that change with every sprint. Developers don't open Confluence; they open their editor. Limn meets them there, adds inline discussion, and makes those docs legible to AI agents via MCP — something Confluence cannot do.

Limn is for developer-owned docs. Confluence is for corporate policy.

4. "We just email each other."

Email produces decisions that are impossible to find six months later. "What was the reason we chose this approach?" becomes a thirty-minute archaeology project. Limn threads are anchored to the exact sentence they discuss, indexed by your repo, readable by Claude in your IDE. The decision and the context stay together forever.

Limn is a permanent record. Email is a temporary channel.

We are betting on a real signal

We are not building features to grow a user base. We are watching one number: do developers who start threads come back the next week? If D7 retention is above 20% and threads-per-user is above 1 per week, Limn is solving a real problem. If not, we will tell you plainly and stop building. We think transparency here is a feature, not a liability.

Start a team trial → How Claude Code reads your threads →