A general AI chatbot doesn't have years of intraday price history for NQ or ES, so when you ask it how often a gap fills, it gives you an estimate that sounds right.
edgeful is a data platform with 150+ reports on how often futures setups have played out, filtered by ticker, session, timeframe, and weekday. It also covers stocks, ETFs, forex, and crypto. In 2026, edgeful added 3 ways to put AI on top of that data: edgeful AI, a custom report builder, and an API with an MCP connector for outside AI tools.
Asking edgeful AI a question
edgeful AI is a chat tool inside the platform. edgeful says it is trained on 8+ years of verified market data and analyzes all 150+ of its reports in real time. You type a question in your own words, and it answers from edgeful's reports.
edgeful AI answering a question about the initial balance breakout on NQ in the NY session, last 6 months (September 22, 2026).
It works best as a research tool. You use it to dig into a setup after the session, ask it questions on how to improve your strategy, or ask it what reports you should be focused on depending on your trader type.
A recent example came from André, edgeful's CEO, who shared it in the company's newsletter. He wanted to build a gap fill setup on NQ, but the full fill numbers looked weak. Over the last 6 months of NY sessions (March 25 to September 24, 2026), NQ gaps filled all the way back to the prior close on 72 of 131 days, or 55.0%.
So he asked edgeful AI this:
"I want to build a gap fill strategy but the stats for the 100% fill are currently pretty bad. What can I do?"
The AI came back with a list of ways to customize the report: a lower fill target, gap size, weekday, the previous candle, fill time, and other tickers. Gap size made the biggest difference. Splitting the same 131 NQ gaps at 0.4% of price gave this:
● Gaps under 0.4% filled all the way 92.9% of the time (52 of 56)
● Gaps of 0.4% or bigger filled all the way 26.7% of the time (20 of 75)
Both numbers come from the same ticker, session, and 6 months, and together they make up the 55.0% above.
A number like that is a starting point. It describes what happened over past sessions, and a trader still has to check the setup on the chart, decide where the stop goes, and test it before putting real money behind it.
Building a custom report from a description
edgeful's 150+ reports cover the common setups: gaps, opening range breakouts, the initial balance, and more. Traders often have a question none of them answers directly, such as how often the low of the day forms between 11:00 and 11:15AM ET on NQ.
The custom report builder, live since August 15, 2026, sits inside edgeful AI as its own build mode. You describe the stat you want in plain English, and edgeful builds the report on the ticker, session, and timeframe you choose.
The flow has a few steps:
● Describe it. Write the question the way you would say it. An "improve prompt" button rewrites a rough idea into a more complete prompt, and it can add details such as a ticker or date range, so it is worth reading before you run it.
● Check the methodology. Before anything is saved, the builder shows what the report measures, every assumption it makes, and a sample results table covering the last 3 months. You can also view the code the report runs on.
● Adjust it. You can ask for changes before saving, such as adding a weekday column, and the builder reruns the calculation.
● Run it. Once saved, the report sits in the analyze mode under custom reports. From there you run it on any ticker, session, and date range.
Custom reports show how often something happened after a given condition. They don't simulate entries, stops, and profit targets, which is what edgeful's automated algos and their optimizer handle.
Pulling the same data into your own AI tools
Some traders already do their research in Claude, ChatGPT, or an AI coding tool. For them, edgeful offers an API, and since August 28, 2026, an MCP connector.
The MCP connector lets an AI assistant such as Claude or ChatGPT call edgeful's reports directly. You add the connector, sign in with your edgeful account, and then ask a plain question like "how often did NQ gaps fill over the last 6 months?" The assistant answers with the number from edgeful's data. AI coding tools such as Claude Code and Cursor use the API directly with an API key.
API access comes with every paid edgeful plan, but the scope changes by plan:
● essential ($49/month): A starter set of reports and tickers, 6 months of history, no live data
● pro ($99/month): All reports and all tickers, live data, and 8 years of history
● all access ($299/month): Everything in pro, plus the automated algos
Pro is the plan built for data-focused traders who want to bring more AI into their process. It opens the full report library to the assistant they already use, so a question in Claude or ChatGPT can run against the same futures trading data that edgeful AI uses inside the platform.
How the 3 fit together
Each tool covers a different kind of question:
● edgeful AI: questions about the existing reports, like the NQ gap fill example above
● Custom reports: questions no existing report answers
● The API and MCP connector: the same questions, asked from Claude, ChatGPT, or your own scripts
All 3 run on edgeful's historical market data.
More on the platform, including the full report list, is at edgeful.com.
This article was written by IL Contributors at investinglive.com.from Investinglive RSS Breaking education Feed https://ift.tt/3IoVstl
via IFTTT
