gex.live Launches MCP Server to Backtest SPX Dealer-Gamma Rules via AI Assistants

·1 views

gex.live has released an MCP server that integrates with AI clients such as Claude, Cursor, and ChatGPT, allowing users to query its SPX dealer-positioning archive and run backtests through natural language. Free tools give access to finished-session data including gamma flip levels, call/put walls, OHLC, and net-gamma percentiles without requiring an API key. A separate set of credit-metered Lab tools enables users to compile plain-language trading ideas into testable rules and run them through a backtesting and optimization pipeline powered by LightGBM. Authentication is handled via bearer token for CLI and Cursor users, while claude.ai uses an OAuth flow with PKCE for header-free environments. The server exposes only historical, finished-session data and not live positioning, and the platform notes that very few tested rules have survived out-of-sample holdout evaluation.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

How to Fix Flutter App Jank: Key Performance Optimization Techniques

Flutter apps can suffer from dropped frames and jank when widget rebuilds, JSON parsing, image decoding, or complex layouts exceed the 16ms frame budget on 60Hz displays. Developers are advised to profile apps using flutter run --profile and Flutter DevTools rather than relying on debug builds, which do not reflect real-world rendering performance. Unnecessary widget rebuilds can be reduced by using const constructors, splitting large widgets into smaller sections, and applying state-management tools like BLoC's buildWhen to limit UI updates. Heavy computations such as large JSON parsing, encryption, or data transformations should be offloaded to background isolates to avoid blocking the UI thread. Additionally, lazy list builders like ListView.builder and careful use of layout constraints can further reduce rendering overhead.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Tutorial: How to Build a Flutter Video Call App Using Agora RTC SDK

A new developer tutorial published on DEV Community walks through building a real-time video calling app in Flutter using the Agora RTC SDK (agora_rtc_engine 6.6.3). The guide covers key steps including initializing the Agora engine, requesting camera and microphone permissions, joining a channel, and rendering both local and remote video streams. It also demonstrates controls for muting audio and toggling video without leaving a call. For production use, the tutorial advises generating short-lived tokens from a secure backend rather than embedding credentials directly in the app. The guide recommends isolating Agora-specific code within a dedicated service layer to keep business logic independent of the SDK.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Guide outlines layered BLoC architecture for scalable Flutter apps in production

A technical tutorial published on DEV Community details an advanced BLoC-based architecture pattern designed for production-grade Flutter applications. The guide emphasizes separating concerns across distinct layers — including presentation, domain, application logic, and data infrastructure — to prevent tight coupling as apps scale. It recommends structuring dependencies so that outer layers depend on stable inner abstractions, keeping business logic independent of specific data sources like Dio, Firebase, or SQLite. The tutorial covers practical implementation of use cases, repository contracts, data models, and explicit sealed state classes to make UI behavior predictable. The approach is intended to help Flutter developers maintain cleaner, more testable codebases as application complexity grows.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Cumora Builds AI Agents as Full Team Members With Coordination Controls

Cumora is an open-source, cross-platform chat application that integrates AI agents alongside human users in direct messages, group chats, a Kanban board, and a calendar. Unlike conventional chat tools, it treats agents as persistent team members with their own identities, memory, email addresses, and the ability to claim tasks. To prevent agents from acting on outdated information or duplicating work, the platform uses a freshness gate that holds stale replies until newer context is reviewed, along with atomic task claims and a triage model. Cumora offers two execution modes: a managed cloud path using the OpenAI Responses API, and a bring-your-own-agent option where users run local models on their own hardware without sending provider keys to the server. The repository documents the architecture in detail but does not provide independent evidence that agent teams outperform single-agent or human-led workflows.

0 comments Read more at DEV Community