Skip to content

linkweave · the self-hosted link-in-bio

One link for everything you make.

Publish a fast page, count every click without tracking anyone, and automate the rest. You own the page, the data and the numbers.

New here? Documentation covers setup, the API and deployment.

Juniper Weaver

Writer and maker. Essays, talks and the occasional newsletter.

  • GitHub
  • Mastodon
  • Bluesky
  • YouTube
  • New essay: weaving the web
  • Newsletter
  • Talks and slides
A sample page with demo data, rendered by the same component as real pages.

Proof, not promises

cookies on public pages
0
IP addresses or user agents stored
0
acceptance rules without a test
0

Every acceptance rule is traced to a test, and pnpm ac:audit fails if one is not.

Why switch

Stop babysitting your link page.

  • Before: A typo goes live the moment you save.

    With linkweave: You edit a draft. Visitors see only the published snapshot.

  • Before: Counting clicks means tracking people.

    With linkweave: Hourly counts per link. No IP address, user agent or cookie.

  • Before: Your site goes stale when your links change.

    With linkweave: A signed webhook on every publish can rebuild it for you.

  • Before: Your page lives on someone else’s terms.

    With linkweave: You run it yourself. You own the page, the data and the numbers.

Features

Everything a creator page needs. Nothing it does not.

  • Publish when you are ready

    Edit a draft as often as you like. Publishing writes an immutable snapshot, so visitors never see a half-finished page. Unpublish at any time.

  • Nothing to leak

    Analytics are hourly counts per link and a coarse referrer. No IP address, user agent or cookie is ever stored.

  • Every click counted

    Each link goes through /r/{id}: a 302 to the target and one click on the counter, with no cookie set.

  • A signed webhook

    One profile.published POST per publish, signed with HMAC-SHA256 and retried with backoff.

  • An Atom feed, free

    Every published page also lives at /your-slug/feed.xml for readers and automations.

  • A typed API with a generated client

    Routes declare zod schemas, the OpenAPI document is generated from them, and CI fails if the committed client goes stale.

  • Locked down

    Zitadel sign-in with PKCE, an owner pin, a Cloudflare edge lock and Postgres row-level security as a second fence.

See every feature, with how-tos

How it works

From zero to published in three steps

  1. Sign in and claim a slug

    One Zitadel login. Pick a slug and it is reserved for you.

  2. Add your links

    Links, social icons, a bio and an avatar. Edit the draft as often as you like.

  3. Publish

    Your page goes live at /your-slug, with an Atom feed at /your-slug/feed.xml.

Walk through it step by step

Questions, answered

Straight answers about privacy, publishing and what linkweave does not do yet.

Does linkweave track my visitors?

No. For each click it stores the link, the UTC hour, a referrer category (direct, social, search or other) and a count. No IP address, user agent, referrer URL, visitor id or cookie is stored, and the public page, the feed and the redirect set no cookies.

Then how do I know which links work?

Every link points at /r/{linkId}, which answers 302 and counts one click. Read daily counts for the last 7 or 30 days, grouped in your time zone, or download the same data as CSV.

What if I publish something by mistake?

Unpublish takes the page offline. The snapshot is kept, so republishing is instant. Visitors only ever see the last published snapshot, never your draft.

Does my page need JavaScript?

No. Public pages are server-rendered, cached and work with JavaScript turned off. Editing in the dashboard does need JavaScript today.

Can I connect it to my own site and tools?

Yes. Each profile can have one signed profile.published webhook that fires on every publish, every page has an Atom feed, and the API is described by an OpenAPI document with a generated, typed client.

Who can sign in?

Only the owners you allow. Sign-in is Zitadel (OIDC with PKCE), and the owner pin admits only the identities you configure: any other account gets 403 on every API route. One owner per deployment is a design choice.

Where does it run?

On infrastructure you control: a Postgres database, a Hono API and a Next.js web app. In production the API answers only requests that come through Cloudflare.

What are the limits?

A rate-limited click still redirects but is not counted (60 per IP hash and 600 per link, per minute). Clicks in flight during a crash or deploy can be lost, and clicks older than 400 days are pruned.

Can I take my data and leave?

Yes. Export your click data as CSV at any time. Deleting your account removes the profile, links, snapshots, clicks and webhook in one transaction.

You will wish you had it yesterday.

Run it locally in five steps, then ship it to your own domain.