ArdaDB
Subsecond SQL on 20 years of tick data.
Every dataset you subscribe to is queryable the moment you sign in. algoseek manages the database, the updates, and the pipelines. You write SQL against 20 years of tick data and get results in subseconds.
| BarDateTime | VolumeWeightPrice | TotalVolume | MaxSpread | TradeAtAsk | TradeAtBid |
|---|---|---|---|---|---|
| 04:00:00 | 130.85412 | 8174 | 0.3 | 1888 | 2079 |
| 04:01:00 | 130.95524 | 8820 | 0.35 | 4774 | 359 |
| 04:02:00 | 131.20877 | 2112 | 0.19 | 78 | 1370 |
| 04:03:00 | 131.22079 | 3888 | 0.09 | 1851 | 123 |
| 04:04:00 | 131.3279 | 5984 | 0.29 | 2442 | 329 |
Tick data at terabyte scale
Databases are easy at megabytes. At terabytes of tick data, you trade speed for cost, or cost for the time you spend maintaining it.
ArdaDB
Setup time
Speed
Cost
Maintenance
Fixed monthly pricing. Your data is already loaded. algoseek manages the database, the updates, and the pipelines. No setup. No surprise bills.
Build your own
Setup time
Speed
Cost
Maintenance
Setup alone takes days or weeks, not hours. But the real cost is what comes after: nightly updates, batch job failures, schema changes, storage growth.
Cloud databases
Setup time
Speed
Cost
Maintenance
Cloud databases charge by compute. One month your bill is $300. The next, you run a batch of historical queries and it is $5,000. You never know what you will pay.
Shared and dedicated ArdaDB
Both handle terabytes of historical research data. The difference is who else is on the server.
Write SQL and go. Your datasets are waiting.
Every subscribed dataset is in the database from day one, updated nightly. Query speed rises and falls with demand on the server, which is what keeps the monthly cost down.
No per-query fees. Fixed compute.
Updated nightly. Pipelines managed by algoseek.
Pricing
Fixed monthly fee
Available with any data subscription. Small egress fees for downloaded results.
Best for: one or two datasets
Your database, your account. Nobody can see your queries.
On Dedicated ArdaDB, information leakage is impossible, as there are no other users. Run production, test, and dev servers, and load your own data alongside algoseek’s.
24/7 monitoring.Dedicated database team.
AWS at cost.No algoseek markup on compute.
Pricing
$500/month
Fixed algoseek service fee. AWS compute costs passed through at cost. Server sizing up to you.
Best for: a full data package
Built for teams that would rather write SQL than manage infrastructure.
Quantitative research teams, fintechs, trading firms without a dedicated database engineering department. You need tick-level market data in a database that works.
A managed market data database
You do not worry about the pipeline, the updates, or the breaks. It is just there. It works.
You never maintain a pipeline
The issue is never the setup. It is nightly updates, batch job failures, schema migrations, and storage growth. algoseek handles all of it, with a dedicated team and 24/7 monitoring.
You never set up or maintain a database
A proper database for terabytes of tick data takes days or weeks to set up, before you write a single query. Everything is already set up and ready to go.
You never worry about scale
Most SQL databases choke on analytical queries across billions of rows of tick data. ArdaDB is built on open-source ClickHouse was built for exactly this, and algoseek extended it for market data.
You never get a surprise bill
Fixed monthly pricing means you follow your research wherever it goes without watching a meter.
Five ways to access ArdaDB
SQL client
Jupyter notebooks
RESTful API
Download to local machine
Python library
Write a query. Get an answer.
The Sandbox gives you up to a year of data, no credit card required. Or talk to the team about a production setup.