Trade Ideas Program API details

Overview

Subscribe to real-time ideas plus consolidated sentiment: both provide current and point-in-time symbology to facilitate mapping and eliminate look-ahead bias.

15 years of history over varying market conditions and 2.6 million+ ideas on 37,000+ stocks globally allow you to train and test models tailored to your specific requirements.

The idea message information includes:

  • Ticker details
  • Direction
  • Author and Author Company
  • Time Horizon
  • Conviction
  • Idea Sizing

The API supports a wide set of mapping identifiers, allowing you to use the data with confidence in your quantitative models.

Details
Coverage
Global
Archive size
Over 2.6m unique ideas
Archive history
Since 2005
Information source
Equities brokers, dealers, IRPs
Content type
Equity trade ideas
Delivery protocol
REST/HTTPS, PULL
Authentication type
Token authentication
Response format
JSON
Update frequency
Real time publishing
Point in time
Yes

Frequently asked questions

Why have we chosen HTTP Long Polling and event sourcing as our distribution API?

Previously this was not the case and our first version was an actual push rather than a logical push. Our learning from this was that most clients preferred to have as much control as possible. Long polling with a consistent event numbering achieves this and gives a good tradeoff between latency and simplicity.

You own the checkpoint, giving you the freedom to replay events at any time, bootstrap new systems without depending on us, and removes the need for reconciliation processes.

You can also upgrade your clients without depending on us to update any configuration – you don’t have to give us access to your systems making them more secure.

How do we ensure that it goes out to all buy-sides at same time?

Open long poll connections are all treated equally, so when a new event is received it is pushed to all open connections. Whilst there can be small differences on individual messages, on average delays will be equally distributed across clients.

What is actually taking place when a contributor puts an idea into the system ?

When an author sends an idea from our Web UI we record the idea event in our event store database. It is then processed by our event pipeline and served to any open long poll connections. Messages are published with the timestamp of when the data was entered into the Web UI.

How do you treat author IDs when two companies merge?

If two companies merge, we move the authors from one company to another so the same author ID will be present for more than one company ID.

What is the uptime/availability of your system?

In the last 6 months we had more than four nines availability outside of scheduled maintenance windows (which are always outside trading hours).

Where is the API hosted?

We run in two DCs in London (Equinix and Digital Realty).