Affitor
Runs affiliate programs for SaaS brands, from click tracking through commission holds to partner payouts.
The problem #
SaaS founders who want an affiliate program choose between enterprise platforms and building the whole thing themselves. The incumbent platforms price for companies that already earn real revenue through partners. A founder with no program yet gets quoted five figures a year before anything is proven.
The approach #
Affitor charged a share of partner-generated revenue instead of a seat fee, so cost tracked results rather than headcount. The first ten thousand dollars of that revenue ran free, which moved the early risk off the brand. That trade only pays when money actually flows through the network, and for Affitor it mostly did not.
Where it is now #
Production runs at api.affitor.com and serves the brands already on it. Fixes ship when a live customer hits something, most recently a Stripe Connect callback bug on 2026-09-12. The ninety day push to win the middle of the market ended in August 2026, and the brand, the domains and the code stay. Affitor LLC is the parent company for later projects, including Kyma API.
What happens if this works #
Attribution should be a signed record both sides can audit, not a cookie owned by one party. If that holds, partner payouts stop depending on trusting whoever controls the tracking script. The verifiable attribution ledger built here is the piece most likely to outlive everything else.
Evidence #
- Live at api.affitor.com and serving the brands on it; last production deploy 2026-08-23 (v3.16.3)
- 37 brand deals worked, 3 reached live status; network fee revenue n/a, pre-revenue
- 7,348 views on the public marketplace, the highest traffic surface in the product
- Six systems in production: tracking API, dashboard, CLI, skills library, public directory, docs
Notes #
Affitor sat in the middle of a two-sided market, and the middle turned out to be the worst seat in the house. The money chain ran brand integrates, partner distributes, customer buys, payout clears, and only then does the network earn its fee. Every link belonged to somebody else, so Affitor could do its own job perfectly and still wait.
The data said it plainly. Three brands signed and reached live status, and the fee revenue was still zero. The strongest demand signal came from brands that had no affiliate program at all, not from brands shopping for a better one; cold outbound to companies already locked into an incumbent converted at a rate that would have needed five times the volume to work.
After ninety days the spending came down rather than running months ahead of the revenue it was chasing. Three things would justify pushing hard again: a different structure, more resources, or a solved cold-start problem. Repeating the same middleman shape is not on the list.
What stayed useful is not the network. It is the attribution architecture that separates fact from policy, the signed ledger demo, and a GTM playbook with real numbers attached to what did and did not work. Those moved on to the next thing.