Dub Open-Sourced the Redirect, Then Sold Attribution

Link vendor marketing still sells click counters, link-in-bio landing pages, and enterprise sales decks from 2010. Dub sells those surfaces too—but the durable bet is quieter. Dub is developer-first link infrastructure: open-source short links, real-time analytics, conversion tracking, and affiliate programs behind one REST API, one webhook stream, and one AGPL codebase you can inspect before you trust a redirect. Judge it on whether your team can programmatically create millions of branded links with sub-150ms latency—not on whether "Dub" clears a generic brand-naming worksheet.

The insight was edge latency and broken attribution, not a category map

Steven Tey built Dub in 2022 as a side project while at Vercel—first to experiment with edge functions and middleware the company was shipping, then to solve an internal pain: sharing links on X, LinkedIn, and iMessage with custom Open Graph previews that actually improved click-through rates (company blog).

The deeper seed was attribution, not vanity URLs. At Vercel, developer advocacy produced articles and tutorials—but revenue impact was opaque. "We didn't know how our devrel work translated into revenue," Tey told TechCrunch. That frustration planted Dub's evolution from "fast shortener" into an end-to-end link attribution engine—first click through signup, purchase, and partner payout.

Dub launched fully open source in September 2022 (introducing post). It hit #1 on Hacker News and trended on GitHub—organic discovery while Tey coded nights and weekends. Over the next year, growth was word-of-mouth with zero paid marketing (company blog). The head of engineering at Twilio discovered Dub through the GitHub repo—openness as enterprise GTM, not a trade-show booth (Mintlify founder interview).

Tey left Vercel in late 2023, incorporated Dub Technologies, Inc., and announced the company in January 2024 (company blog). The product rebranded from dub.sh to dub.co in August 2023—a .co home for link infrastructure, not a .ly hack on someone else's ccTLD.

Open-core model: AGPL inspection, cloud scale

Dub's enduring strategy is not "be a prettier Bitly." It is two compounding layers with a license boundary:

| Layer | What Dub open-sources | What Dub sells | | --- | --- | --- | | Core platform | ~99% of the repo under AGPLv3 on GitHub—short links, analytics, dashboards | Trust, fork rights, community contributions (24K+ stars, 90+ contributors) | | Enterprise Edition | /ee folder under commercial license—SSO, audit logs, enterprise-only features | Larger orgs that need governance without forking | | Managed cloud | Same codebase you can self-host | Edge latency, ClickHouse/Tinybird analytics, PlanetScale/MySQL, Upstash queues—ops Dub runs | | Partners product | Partner program UI patterns in OSS | Automated payouts, tax compliance, embedded referral dashboards, migration from Rewardful |

Dub Technologies calls itself a commercial open-source company—the AGPL core is real, not a demo repo with the API paywalled. OSS Capital led a $2M founding seed in January 2025 precisely because links plus attribution is a massive addressable market and open source is distribution (TechCrunch). Angels include Guillermo Rauch (Vercel), Balaji Srinivasan, Framer founders Jorn van Dijk and Koen Bok, designer Soleio, and Cal.com chairman Peer Richelsen—founder-operator syndicate from the Vercel/design-tool orbit, not a generic SaaS fund.

Note what verified filings do not show: Dub has no Y Combinator company page and its disclosed round is OSS Capital seed, not a YC batch. The gravity is YC-adjacent stack energy (Vercel, Framer, Cal.com)—price the product on API depth, not logo bingo.

Product arc: short links → attribution → partner infrastructure

| Era | What shipped | Strategic move | | --- | --- | --- | | Sep 2022 | Dub open source; free custom domains + analytics | OSS wedge; HN/GitHub distribution | | Aug 2023 | Rebrand dub.sh → dub.co | Infrastructure brand on .co, not ccTLD rent | | Jan 2024 | Company announcement; CEO full-time | Side project → Dub Technologies, Inc. | | 2024 | REST API, TypeScript/Python/Go/Ruby SDKs, webhooks | Developer-first vs marketer drag-and-drop | | Jan 2025 | $2M OSS Capital seed; Conversions product | Click → signup/sale; Stripe/Shopify integrations | | 2025–2026 | Dub Partners scale; Rewardful migrations; MCP server | Affiliate/referral as second product surface |

Dub now positions as three pillars on one platform: Dub Links (branded short links, QR, geo/device targeting, A/B tests, deep links), Dub Analytics (real-time events, conversion tracking, customer insights), and Dub Partners (embedded affiliate programs, global payouts, partner network). Public metrics from the README: 100M+ clicks and 2M+ links monthly; customers include Twilio, Buffer, Framer, Perplexity, Vercel, Laravel, and more (customers). Mintlify reported 1,000+ paying customers without paid marketing—classic open-source PLG (interview).

The architectural thesis is edge-native redirects close to users—sub-150ms request latency at production scale per customer quotes on dub.co/links—with analytics piped through ClickHouse and Tinybird, not a 2010 PHP redirect farm.

Integration starts with one API call—not a marketing UI session:

import { Dub } from 'dub';

const dub = new Dub({ token: process.env.DUB_API_KEY });

const link = await dub.links.create({
  url: 'https://acme.com/pricing?utm_source=newsletter',
  domain: 'go.acme.com',
  key: 'spring-sale',
  trackConversion: true,
});
// link.shortLink → https://go.acme.com/spring-sale

Webhooks fire on link.created, link.clicked, and conversion events—Bitly charges enterprise tiers for API access many teams never get on free plans; Dub ships API + webhooks on free tier (pricing). That is the developer bet: links become programmable infrastructure in your CI, CRM, and product—not rows in a marketer's spreadsheet.

Business model: generous free tier, attribution upsell

Dub monetizes through Dub Links and Dub Partners plans that scale with links, tracked events, and partner payouts (Links pricing, Partners pricing).

Representative Dub Links tiers:

| Plan | Price | Links / events | Sweet spot | | --- | --- | --- | --- | | Free | $0 | 25 new links/mo, 1K tracked events/mo, 3 custom domains, 30-day analytics | Indie devs, OSS users, API experiments | | Pro | $30/mo | 1K links/mo, 50K events/mo, 1-year retention | Creators needing folders + deep links | | Business | $90/mo | 10K links/mo, 250K events/mo, conversion tracking, A/B tests, webhooks | Startups replacing Bitly + Rewardful | | Advanced | $300/mo | 50K links/mo, 1M events/mo, elevated API limits | High-volume programmatic link factories |

The free tier is deliberately aggressive: custom domains, full analytics, and API access where Bitly's free plan offers 5 links and no custom domain (competitive comparisons). Dub wins PLG when engineers install the SDK before procurement notices.

Dub Partners adds payout rails—$90/mo Business includes $2.5K/month partner payouts; Advanced scales to $15K/month with embedded dashboards, tax compliance, and migration tooling (Rewardful → Dub in "2–3 clicks" per Tey's 2026 posts). Revenue compounds when marketing buys links and growth buys affiliate infrastructure on the same event stream.

Competitors: same redirect, different default buyer

HTTP 302 is the wire format; the custody and attribution contract is not interchangeable.

| | Dub | Bitly | Short.io | Rebrandly | | --- | --- | --- | --- | --- | | Built for | Dev-led teams needing API + attribution + partners | Enterprise marketing + link-in-bio brand recognition | Budget-conscious branded links at volume | Agencies prioritizing branded URL consistency | | Open source | AGPL core (~99%) | Closed | Closed | Closed | | Free tier | 25 links/mo, 3 domains, API + analytics | 5 links/mo, no custom domain on free | 1,000 links/mo, 5 domains (generous volume) | 10 links/mo, 1 domain | | API on free | Yes | No (paid tiers) | Limited | Limited | | Webhooks | Yes (Business+) | Effectively no | Paid tiers | Limited | | Attribution depth | Click → conversion → partner payout | Mature analytics; weaker dev-native attribution story | Strong routing (geo/device); less partner infra | Brand controls; fewer programmatic primitives | | Sweet spot | Programmatic links + OSS inspectability + affiliate expansion | Enterprise SLA, link-in-bio, QR customization | High free-tier link volume | Marketing-team branded links without building infra | | Weak spot | Younger enterprise track record vs Bitly | Dated DX; API paywall; expensive advanced features | Not built as attribution + affiliate platform | Fewer native dev integrations (Rebrandly cites ~8 vs Zapier-heavy stacks) |

Choose Dub when engineers own link creation in code, you want webhooks + conversion tracking + partner programs on one vendor, and AGPL matters for trust or self-host evaluation.

Choose Bitly when brand recognition, link-in-bio pages, and enterprise procurement comfort beat API economics—and budget is not primary.

Choose Short.io when you need maximum free branded links (1,000/mo) and geo/device routing without affiliate complexity—classic link utility, not attribution platform.

Choose Rebrandly when marketing teams want branded links + QR with agency workflows—not a GitHub-first OSS story.

Self-host Dub when AGPL custody is mandatory and you can staff PlanetScale, ClickHouse, Tinybird, and edge ops—accept that enterprise /ee features and managed payout rails stay on cloud.

China: global link infra vs mainland short-link politics

Dub's China story is not "Dub 中国." It is Chinese startups standardizing on the same Vercel-adjacent dev stack for global products—a different problem from owning mainland redirect infrastructure.

What Dub does well (出海 / global SaaS):

• Chinese AI tools, dev products, and indie SaaS targeting US/EU users often already run Dub + Vercel + Stripe + Linear—the same gravity that made Guillermo Rauch an angel check rational.
• Programmatic link APIs let distributed eng teams generate campaign URLs from CI without a marketer's Bitly login.
• Conversion tracking with Stripe/Shopify integrations suits global checkout flows—not domestic super-app wallets.
• Open-source AGPL appeals to teams that want to inspect redirect logic before sending production traffic through a third party.
• Dub works with Malaysia's government on open-source link infrastructure per TechCrunch—signal that OSS deployment models translate to regulated public-sector use cases, not only SF startups.

What Dub does not replace (境内链路):

• Mainland-facing apps often use ** domestic short-link and analytics stacks**—微信短链、字节系投放链接、百度统计、友盟——not dub.co endpoints that may face latency, blocking, or ICP-adjacent scrutiny on custom domains.
• Custom domains for Chinese commercial sites require ICP备案 and local DNS relationships Dub does not provide.
• Affiliate and creator attribution on domestic platforms (抖音、小红书、淘宝联盟) lives inside closed ecosystems—Dub's partner payouts do not replace 巨量引擎 or platform-native CPS.
• Deep links to super-app WebViews follow WeChat/Apple universal link rules Dub supports generically but does not localize for PRC compliance reviews.

Practical split-stack for Chinese teams:

| Audience | Typical stack | Why | | --- | --- | --- | | Global product (.com) | Dub Cloud + custom domain | API, webhooks, conversion + partner infra | | Mainland consumer app | 平台原生短链 + domestic analytics | Local compliance + ad network feedback | | Global HQ + China eng | Dub for outbound campaigns; domestic tools for .cn | Two redirect policies—two privacy disclosures | | AGPL self-host | Fork Dub on your cloud | You patch; you operate ClickHouse + edge |

The old startup-showcase template implied one link vendor "works everywhere." Reality: Dub wins programmatic attribution for global dev stacks; mainland campaign links still reward domestic platform gravity.

The name and domain (briefly—facts beat folklore)

The verified story is shorter than a naming-pillar worksheet:

• "Dub" is short, typeable, and memorable—Tey's side-project name that survived from dub.sh (September 2022 launch) through the August 2023 rebrand to dub.co. It evokes "dubbing" a URL onto a shorter alias without locking the brand to one TLD hack.
• dub.co is product, docs, dashboard, and pricing on a clean .co—infrastructure signaling like Stripe on .com, not rent-seeking on .ly or .to.
• dub.sh remains the default short domain in examples (dub.sh/about-dub)—separate marketing hostname from customer custom domains (go.acme.com) and complimentary .link domains on paid plans.
• The GitHub org moved to dubinc/dub—OSS brand federated to the commercial API, same pattern as React Email → Resend.

That is engineer-native naming—three letters, one syllable—not a five-act "domain scorecard."

Prove the infrastructure: a link custody drill

Developer trust is not a homepage adjective. Run this before you treat "open source means portable" as full exit freedom:

1. Separate redirect custody from analytics custody. AGPL source survives a fork. List what does not: Dub Partners payout history, embedded referral configs, dashboard-only folders, and Partner Network relationships.

2. Export links and event schemas. Pull links via API; replay webhook payloads (link.clicked, conversion events) against your warehouse. Confirm UTM templates and A/B test arms map cleanly.

3. Latency baseline. Create 1,000 links via bulk API; measure redirect p95 from US/EU/APAC vantage points. Sub-150ms is Dub's production claim—verify on your domains, not dub.sh alone.

4. Self-host smoke test. Deploy dubinc/dub per self-hosting docs with PlanetScale + ClickHouse + Tinybird. One tracked click end-to-end proves AGPL custody—not a commitment to run link infra forever.

5. Partner migration rehearsal. If you run Rewardful or a spreadsheet affiliate program, test Dub's migration tool on staging partners before promising "2–3 clicks" to finance.

What to verify before your team standardizes on it

1. Links vs Partners custody. Short links and affiliate payouts are different compliance regimes—confirm tax forms, payout limits, and 5% payout fees match your finance model (Partners pricing).

2. Free-tier math. 25 new links/month is tight for high-velocity campaigns; Short.io's 1,000 free links may win volume-only bake-offs even if API attribution loses.

3. Bitly incumbency. Marketing may still want Bitly link-in-bio or QR customization—count SKUs you'd retire, not SKUs you'd add "later."

4. China audience split. Mainland campaign links likely need domestic tools regardless of how clean Dub's API docs read in English.

5. OSS vs cloud coupling. You can fork AGPL core; you cannot fork managed partner payouts or Enterprise /ee without a migration project—decide which layer is system-of-record.

Dub's competitive edge is not the loudest slide in a "links are back" webinar. It is open-source link infrastructure fused to attribution and partner programs—$2M OSS Capital seed, 24K+ GitHub stars, 100M+ monthly clicks, customers from Twilio to Perplexity, and a bet that the next trillion links will be created by APIs, not dashboards. Price the redirect—and run the custody drill before your stack becomes "Dub plus the Bitly login marketing secretly kept."