Glitch Executor · Holding company
AI systems that move
money, prices, and signals.
One agent stack, three product lines. Glitch Trade executes in markets, Glitch Edge reads the field, and Glitch Grow runs e-commerce — sharing data, models, and battle-tested infrastructure across all three.
- Product lines
- 3
- Ensemble models
- 8+
- Shared stack
- 1
Portfolio
Three product lines. One operating system.
Each sub-brand runs on the same shared stack — agent ensemble, shared data layer, shared execution primitives — but ships with a vertical-specific surface area. Pick the one closest to your problem.
-
Glitch Trade
trade.glitchexecutor.com
AI trading ensemble across cTrader, MT5, and IB.
- 8-model ensemble — trend, momentum, mean-reversion, ML, volume, session, MTF, technicals
- cTrader Open API, MT5, Interactive Brokers — one execution layer
- Telegram client bot with signal alerts, RAG assistant, billing
-
Glitch Edge
edge.glitchexecutor.com
Sports intelligence engines for professional bettors.
- Pre-match and in-play model stack with edge-aware staking
- Private data pipelines — no scraped-odds dependency
- Built for high-volume syndicates, not retail tipsters
-
Glitch Grow
grow.glitchexecutor.com
AI digital marketing for DTC and e-commerce.
- Cross-platform attribution — Meta, Amazon, Shopify unified
- Voice AI for COD confirmation — training data you own
- Shopify automation — theme-as-code, multi-store safe
How we work
Same agent patterns, different surface areas.
Trade, Edge, and Grow look like three unrelated products from the outside. Under the hood, they share the same router, the same data contracts, the same execution primitives — so every improvement compounds across the portfolio instead of staying siloed.
-
One ensemble, many domains
The same 8-model router that ranks trade signals also scores betting edges and ad-spend allocations. Trend, momentum, mean-reversion, session analysts — the patterns transfer once you own the plumbing.
-
Owned data, not rented access
Every engagement leaves the client with their own Postgres, their own embeddings, their own audio and call datasets. Fine-tunes run on your data inside your infra — we do not recycle it into ours.
-
Cross-domain pattern reuse
The attribution bridge built for a DTC brand informed how our trading executor filters noisy signal feeds. Work done in sports modelling sharpens marketing conversion windows. One codebase, three revenue surfaces.
-
Vendor-lock-in avoidance, on purpose
Open-weight models where latency and privacy matter. Frontier models where reasoning matters. cTrader, MT5, Interactive Brokers all pluggable behind one executor interface. No single API deprecation ends the product.
-
Edge compute for the thin bits
Webhooks, OAuth callbacks, contact forms, OG image generation — Cloudflare Workers and Pages Functions. Heavy ensemble work runs on dedicated GCP. Split by where latency hurts most.
-
Code is public by default
Where it is safe to do so, our internal libraries ship under a BSL 1.1 licence that converts to Apache 2.0 in 2030. Transparent moat: the moat is the data and the ops, not the source.
Selected milestones
A short backstory.
The company runs lean: one small team, one shared server footprint, three product lines shipping against real clients. Milestones below are the ones we can talk about publicly.
- 2026-04
Four-site portfolio rebuild
Master portfolio and three sister marketing sites (Trade, Edge, Grow) rebuilt on a shared Astro + Cloudflare Pages stack. Shared cyber-cobra mascot and design language across the estate.
- 2026-04
Ayurpet — multi-storefront attribution live
Cross-currency, cross-storefront attribution in production for an INR ad account driving both India and Global Shopify stores, with Amazon Seller Central sync via Airbyte → Postgres → agent view.
- 2026-04
Public-repo licensing standard adopted
BSL 1.1 → Apache 2.0 (2030-04-18) adopted across public agent repos — source-available now, fully open-source on schedule. Transparent moat: data and ops, not source code.
- 2026-04
cTrader Open API integration
cTrader Open API client and price feed built and registered with Spotware for live execution across forex, commodities, and equities, joining existing MT5 and Interactive Brokers paths.
- 2026-04
Voice AI for COD confirmation
Production voice agent handling Hindi and Hinglish COD confirmation calls for Shopify merchants. Every call becomes training data the merchant owns — not a moat that expires when we leave.
- Ongoing
8-model trading ensemble
Trend, momentum, mean-reversion, ML, volume-profile, session, multi-timeframe, and technicals — routed through a shared ensemble engine feeding the Telegram client bot, admin API, and execution workers.
FAQ
The questions we get on the holding-co side.
Product-specific questions live on each sub-brand site. These are the ones that only make sense at the portfolio level.
-
What does Glitch Executor actually sell?
Three product lines under one brand. Glitch Trade is an AI trading ensemble (cTrader, MT5, Interactive Brokers) sold to retail and pro traders via a Telegram client and admin dashboard. Glitch Edge is a sports intelligence engine for professional bettors and syndicates. Glitch Grow is an AI digital-marketing agency for DTC and e-commerce brands — Meta, Amazon, Shopify, Voice AI. Each has its own site (trade. / edge. / grow.) and its own go-to-market; this page is the holding-company portfolio. -
Are these productized or bespoke?
A mix. Glitch Trade is the most productized — signed-up clients get a working Telegram bot on day one. Glitch Edge is currently delivered to a small number of design partners and is not open to public sign-ups. Glitch Grow runs a fixed-scope 30-day pilot followed by month-to-month engagement, so it's productized at the pilot boundary and bespoke inside it. If you need a clean line between "subscribe" and "hire us", ask on the first call. -
Do the three products share tech?
Yes, intentionally. One ensemble router, one Postgres + pgvector data layer, one edge-compute footprint on Cloudflare Workers, and one shared execution-adapter pattern. The 8-model trading ensemble and the betting engine share most of their scoring primitives; attribution work built for e-commerce informs how the trading executor filters noisy signal feeds. Improvements compound across the portfolio instead of staying siloed per-product. -
Are you hiring? Open-source?
Not actively hiring — the team stays deliberately small until each product line has its own revenue to fund a dedicated ops headcount. Public agent repos ship under BSL 1.1 today and convert to Apache 2.0 on 2030-04-18. You can read, fork, and self-host now; you can build a competing hosted product on it starting in 2030. If you want to contribute, open an issue before a PR — we optimize for narrow scope. -
How is the company structured?
Glitch Executor Labs is the parent entity; Glitch Trade, Glitch Edge, and Glitch Grow are product lines underneath it, not separate companies. All three currently run on shared infrastructure; we split into isolated VMs once any one product line reaches meaningful standalone revenue. Security blast-radius is the trigger for that split, not arbitrary head-count. -
How do I reach the team?
Email support@glitchexecutor.com and tell us which product line fits your problem. If you're not sure, describe the outcome you need — we'll route you to Trade, Edge, or Grow (or tell you honestly if none of them is the right fit yet). We reply within one business day.
Get in touch
Tell us which product line fits your problem.
We reply within one business day. If it's urgent, email us directly — support@glitchexecutor.com.
- Trading, betting, or marketing — we'll route you to the right sub-brand.
- No NDAs required to scope the first call.
- Honest answers if none of the product lines is the right fit yet.