Treg - OpenRouter for data & tooling
Agents need the outside world: people and company data, search, social, SEO. treg gives them one key to a catalog of external APIs and bills per call. Since mid-September, calls have grown more than fourfold to over 2M a day and the revenue run-rate has grown from $0.66M to $2.1M.
The biggest marketplace for agent to discover, access data & tools
The same playbook as Costco: win on price and selection first, learn more about demand than anyone, then sell your own products where the market is weak.
Aggregate
One account for every tool an agent needs, at vendor price with 0% markup. Customers have no reason to go direct, so volume concentrates on treg, and vendors discount it.
Today: 40M+ requestsBenchmark
Every call shows which provider answered, how accurately, how fast and at what cost. No vendor can see this across its competitors. It becomes routing, so agents pick the best tool per request, including tools nobody integrated.
Today: launched Enrich Arena as a first tryOwn the gaps
The benchmark shows where vendors are expensive, inaccurate or missing. We fill those gaps with first-party data and tools, our Kirkland Signature, sold into demand we already see.
Today: our team has deep expertise here ↓0% markup to customers, always. Margin comes from vendor volume discounts, which grow with scale, and from first-party tools we build.
More usage & volume → better bulk discounts → more benchmark data & insights for first-party tools → more usage.
8 weeks from 0 to $2M+ run-rate
Seven-day average of daily revenue, annualised. Weekly revenue nearly tripled in three weeks after the launch-week spike rolled off: +49%, +69%, then +12% in the last full week, to $35.4k. The seven days to 7 Oct reached $40.6k, a $2.1M run-rate, including a record $9.8k day.
Annualised run-rate (USD), 7-day trailing average
2M+ daily requests, 40M+ cumulated
Daily calls went from under 10k in mid-August to 3.09M on 7 Oct. The 1 Sep launch lifted usage to about 600k a day, and it has grown about fivefold since. Calls rose 48% in the last seven days.
Calls per day
Full PLG motion, with no sales team
Every company here found treg on its own, signed up and paid with its own card. There were no sales calls, pilots or contracts. The people using it are the ones who own growth: sales teams, GTM engineers and founders, including the CEOs of Appsmith, Relevance AI and Wordware.
+ 2,787 more paying teams
Company figures are the latest publicly reported; older valuations show their year. Calls are all-time through 28 Sep.
GTM: Horizontal marketplace, vertical launches
Instead of one big launch, we ship treg into a new go-to-market job every week or so. Each post walks through a full workflow, with the price per call and the open-source code behind it, so it doubles as a tutorial. Most people first find treg this way.
Eight more launches have followed, from Enrich Arena to a Clay rebuild for agents. Together, posts about treg have reached 3.9M views on X since 11 Aug.
Community organic sharing
Unprompted posts from people showing the workflows they built on treg.
1 in 7 sign-ups pays, most on day one
Signups are people. Everything after that is counted per team, because a team holds the balance and makes the calls. 39% of all signups are teams still calling ten or more times a month.
Signup to payment, by stage
64% of signups reach a real call, and 56% of paying teams paid within 24 hours of signing up (median 14 hours).
60% revenue from repeat purchases
Returning payers brought in 60% of last week's revenue. Auto-recharge, where a team's balance refills itself, more than doubled its share of revenue in three weeks, to 24%.
Weekly revenue, first purchase vs repeat
Is the revenue whale-based?
No. It is a normal usage power law with no single-customer risk. The largest customer is 2.8% of revenue, and the top tenth of payers carries about half. That is the shape healthy usage-based pricing produces, not a handful of accounts holding the company up.
Share of revenue held by the top N% of payers
The bottom half of payers still contributes 11% of revenue, and the top 10 customers together hold 11.9%.
Land with one job, expand into many
Most teams come for one job, usually finding leads, and 47% now use two or more. Every empty search and tool request shows us what agents want next. When we add it, existing customers adopt it within weeks.
Where teams start
First job each team used treg for, across the 11,581 teams with 10+ calls.
25% of these teams use three or more jobs.
The data flywheel
The flywheel at workTwo jobs we added because agents kept asking for them, and what happened next:
Calls and jobs exclude refused requests. *Week of 5 Oct is Monday to Wednesday. Highlighted logos are vendors added after the demand appeared.
Competitors: ~4× the leading aggregator, neutral above vertical SaaS
Two groups compete for the same agent spend. Among other aggregators, treg carries about 4× the daily traffic of Monid, the only one that publishes its numbers. Single-vendor SaaS tools become suppliers on treg, not substitutes for it.
1 · Agent data and tool aggregators
Same model as treg: one key, one balance, many APIs. Competitor figures are their own public claims.
| Company | Traffic | Backing |
|---|---|---|
| treg | 2M+ requests/day | — |
 Monid | ~0.5M/day (their team, Discord, Oct)our estimate from their public chart: ~12M in Sep vs treg 25M | $7.7M raised |
 Orthogonal | Not disclosed | $4.3M seed (Pantera), YC |
 AIsa | Not disclosed | $6.5M (Alibaba, Tribe) |
 Locus | Not disclosed | YC F25 |
2 · Vertical SaaS: Clay, Apollo, Ahrefs, Semrush, Profound
Each of them is one vendor. treg aggregates all the vendors in a category and learns from every call which one is best, so agents get better results at a fraction of the cost.
Vertical SaaS
Clay
ApolloClay resells a sales-data waterfall at a credit markup; Apollo sells its own database.
treg · 40 vendors integrated
ContactOut
LeadMagic
Findymail
Hunter
Prospeo
Lusha
People Data Labs
Crustdata
Coresignal
Fiber AI
AI Ark
Wiza
Icypeas
Tomba
Datagma
Aviato
EnrichLayer
Limadata
Leadsforge
QuickEnrich
Kitt
Dropleads
GetLeads
HarvestAPI
Moltsets
Sumble
PredictLeads
Crunchbase
CompanyEnrich
The Companies API
Ocean.io
OpenmartDiffbot
Brand.dev
Akta
Trestle
Scrubby
MillionVerifier
ZeroBounce
BounceBanEvery call is scored, so each lookup goes to the vendor that is most accurate and cheapest for that input.
Most correct emails, at 85% less than Clay
Same leads sent to every tool, name and domain in, graded against the address each person publishes (100 leads, Oct 2026).
Correct email found, of 100
Clay84
Hunter84
Deepline80Apollo77Cost per correct email
Hunter$0.0263Apollo$0.0290Deepline$0.0340Clay$0.0377Then on 100 random real customer requests, each tool given exactly what the customer sent:
Deliverable emails found, of 100
Apollo24Clay20Deepline13Phone numbers found, of 30
RocketReach14Apollo13Clay13Competitor costs at list price on the day. Differences of a few rows are within noise; the cost gap is not.
The same playbook in every other category:
Vertical SaaS
Ahrefs
SemrushEach sells its own index behind a monthly plan.
treg · 10 vendors integrated
Semrush
DataForSEO
Moz
SE Ranking
Serpstat
SpyFu
Majestic
SerpApiSerper
Search ConsoleSemrush is one of our vendors. Agents pay per call and get whichever index answers best.
Vertical SaaS
ProfoundOne dashboard, sold by annual contract.
treg · 4 vendors integrated
cloroDataForSEOSerpApi
JustOneAPIRaw answers from ChatGPT, Perplexity, Gemini and Google AI Overviews, per call.
The most founder-market-fit team
A founder who grows developer products for a living, and engineers who built the tools agents already depend on. treg sits at the edge of what the most advanced builders use, and it grows through product-led adoption. That is exactly this team's skill set.
Hyper-PLG growth for AI builders is his craft. His audience is the builders treg sells to.
Firecrawl
TinyFish
DaytonaOrcaHead of Product & AI, Relevance AI
One of the best web-scraping engineers anywhere. Leads first-party data for step 3 and keeps treg credible in open source.
ByteDanceLed AIGC product at ByteDance. Knows what agents need from tools because he built one of the most-used agent frameworks.
AlipayByteDanceSolved Alipay's payment-scale concurrency. treg's hardest scaling problem is the same one: billions of metered calls, each with money attached.
aiseo.com · 1M+ users at peakHas already scaled an AI product to a million users.