You’ve heard of data vendor lock-in, but
Have You Heard of Data Vendor C*ck Blockin’?
Every data vendor started as one person who understood a source, then grew an org chart around them. Now that an agent builds the dashboard in an afternoon, the only thing worth paying for is the person who keeps the feed correct. Dryos pays them, per query, and nobody else.
- no credit card required
- clean, validated data
- beautiful visualizations
ERCOT prices, ancillary services and load. Knows which report drops its repeated-hour flag.
Dryos is a marketplace.
We work with maintainers who know a data source better than anyone else. They get paid for collecting and maintaining it: adapting when the source changes, making revisions when something needs cleaning up, and getting it to Dryos fast.
A maintainer claims a source
Someone who knows the data source better than anyone else. They collect it, adapt when the source changes, make revisions when something needs cleaning up, and get it to Dryos fast.
Dryos hosts, validates and meters it
Every run is checked against the source before it is promoted. Every query passes one place that counts it. That is the whole of what we add.
You use the data
In a workspace, through the API or the MCP server, or delivered into the Snowflake or Databricks warehouse you already treat as the truth. A stream costs the same whichever way it arrives.
They get paid
When you use the data, the maintainer earns. That's it. No org chart.
Your usage flows back to the maintainer. Not to a sales quota, not to a lease, not to a roadmap.
The data is public. The cleanup is the job.
Almost everything worth building on is already published: by a grid operator, a weather service, a county. It arrives as zipped CSVs, hourly republications, local-time timestamps and column names that change without notice. Formatting, normalising and cleaning it is the whole job. That is what every data vendor charges for, and each of them does it for one vertical only.
Energy
30 live streams- Pricing
- Load
- Generation
- Ancillary
- Grid
ERCOT market reports and grid dashboards
Zipped CSVs on a listing that keeps a few days. Timestamps in Central time, with a repeated-hour flag every report carries except one. Forecasts republished hourly, each copy a new vintage of tomorrow.
Sold today by vendors who do only energy, to companies who need only energy.
Weather
4 live streams- Observations
- Forecast
National Weather Service observations and forecasts, Open-Meteo wind fields
A pagination cursor that never says stop. A field the station did not report and a field that does not exist, both arriving as null. Units that change from one station to the next.
Sold today by vendors who do only weather, to companies who need only weather.
Property
next- Assessments
- Transfers
- Permits
- Zoning
County appraisal districts, clerks and permitting offices
Published county by county, each in its own schema: some as spreadsheets, some as PDFs, none as an API. The same parcel described three ways by three offices.
Sold today by vendors who do only property, to companies who need only property.
Every vendor cleans one vertical. We clean all of them, once.
A single-vertical vendor builds a collector, a validator, hosting, billing and a sales team, and charges one industry for the lot. One vertical over, another company builds the same five things again. The data was public both times. What got paid for twice was the plumbing.
Dryos is horizontal, so the plumbing is built once. The pipeline that stages, validates and promotes an ERCOT report is the one a weather station goes through and the one a county roll will; the sandbox, the metering and the marketplace are shared by every stream. The fixed cost is spread across every vertical, and what a stream has to earn is its maintainer’s time and a slice of infrastructure, not a company. That is a different unit economics, and it is why the invoice below has two lines.
Nobody should be paying for an org chart.
A data vendor’s invoice pays for the whole company that grew around the data. The sales quota, the conference booth, the lease, the executives, and yes, the birthday present. All of it lands on your bill. A Dryos invoice has two lines: our fee, which covers infrastructure, databases and hosting, and the maintainer’s time and expertise, which they are rewarded for fairly.
- Sales team26%Whose quarter depends on you renewing.
- Customer success16%Onboarding decks and quarterly check-ins.
- Marketing and conferences15%The booth, the lanyards, the steak dinner.
- Executive compensation18%
- The CEO's daughter's 16th-birthday Porsche2%Happy birthday. You paid for it.
- Office lease9%
- Legal and procurement6%Six weeks of contract for a two-day integration.
- The person who actually understands the source8%
- Dryos fee30%Infrastructure, databases, hosting, validation and metering.
- The maintainer's time and expertise70%Paid every time you query. The feed being right is their whole job.
Illustrative shares, not anyone’s audited books. The Porsche is a guess. The proportions are not.
Build on what your team already built.
AI made a dashboard an afternoon’s work, so now everyone builds their own: the same feed, the same chart, four times over, none of them shared. A workspace is where that stops. Pages are put together from components somebody already made, a wired group lands as one drop, and every change is recorded as the sentence that asked for it. Do not repeat yourself, as an operating system.
Six tiles on this afternoon’s ERCOT prices and none of them coded. The map, the chart and the ticker are one published group, already wired to each other; load by zone and the fuel mix are published components; the heatmap is a shape dropped onto a stream. Captured from the running page, on the live feed.
Wire it once. Everyone drops the group.
Click a node on the map and the chart and the ticker follow it. Somebody wires that once and publishes it, and it lands on every other page as one drop, already connected. The wire is stored on the receiver and replays like any other setting, so it survives the copy.
On the launched screen both ends wear the pair’s colour, so whoever inherits the page can see what is driving what without asking who built it.
Bring your own model key. Claude, OpenAI, whichever you already pay for, on hosted and self-hosted alike. We do not mark it up; that bill stays between you and them. AI mode gets the full model, and what it lands on is a typed component: generated code that compiles or is refused, so the page your team builds this quarter still runs a year from now.
Flat per month. Streams, not queries.
A stream is a thing somebody maintains, so it is the thing you subscribe to. The platform is a flat fee. Nothing on the invoice depends on how many times a tile polled, which means you can budget it and we can forecast it.
Free
One workspace on the public feeds, to see whether the data is what you think it is.
- 1 workspace, 3 streams
- Standard polling
- Every shape, the community shelf
Team
A desk's worth of screens for a team that shares them, on the streams it actually uses.
- 5 seats, 10 streams included
- $15 a month per extra stream
- API and MCP within a fair-use quota
- Overage credits at list, hard cap by default
- Bring your own model key, no markup
Business
Shared workspaces across the company, and the streams delivered into the warehouse you already trust.
- 20 seats, 30 streams included
- SSO and shared workspaces
- Delivery into Snowflake or Databricks
- Priority support and an account manager
- Annual commit takes 15 to 20% off
Self-hosted
Fork the repository and run the platform yourself. Queries are yours, so they are unlimited; what you subscribe to is the data.
- Minimum 10 streams
- Updates and support included
- Your own model keys, no markup
- The MCP server, for you and your AI code companion
Consultant
Build workspaces for your clients on a Team plan each. You see their usage and add a percentage or a flat fee for your work.
- Save components and pages, reuse them across engagements
- Billing managed in the app
- White labelled for your brand
Flat, per month
The fee and the streams are the invoice. No line on it moves because a screen was left on over the weekend.
Streams, not queries
A stream costs the same whether one screen watches it or fifty. Overages exist only past a fair-use quota, opt-in, capped by default, with a budget alert first.
Commit or prepay
An annual commit takes 15 to 20% off. Prepaid credit packs, valid twelve months, for the procurement team that wants one number.
70% to the maintainer
Of every stream's subscription. The platform fee covers the plumbing; the data line pays the person.
Do not repeat yourself operating system
We are Dryos. Built on the belief that in the new AI landscape, data vendors should meet their customers where they are.
We believe in providing a great service, not in locking any company into an agreement. Data maintainers should earn based on the value they provide. Your organization deserves a different kind of data company.
Only build what you need, and do not repeat yourself.