Technology Stack
Technology Stack
Built for Reliability
Ascension's infrastructure is designed around a simple principle: the system should run without constant attention. That means choosing proven technologies, building in redundancy, and keeping the architecture clean enough that problems are easy to diagnose and fix.
We're not chasing cutting-edge for its own sake. We're building something that works.
Infrastructure Layer
Cloud Hosting The system runs on cloud infrastructure with automated process management. Services restart if they fail. Logs capture what happened. The goal is uptime without babysitting.
Containerization (Phase 2+) As the system scales to support multiple bot instances for renters, containerized deployments will allow isolated, reproducible environments for each user's configuration.
Trading Layer
Language & Framework The trading bots and supporting systems are built in Python — chosen for its extensive library ecosystem, rapid development cycle, and strong support for financial data processing.
Exchange Connectivity Bots connect to exchanges through official APIs and established wrapper libraries. The architecture supports multiple exchanges and can expand to new ones as needed.
Data Storage Trade logs, performance metrics, and system state are stored in lightweight databases optimized for the read/write patterns of trading systems. Nothing fancy — just reliable.
Intelligence Layer
Market Data Real-time and historical price data flows from exchange APIs and third-party data providers. The system ingests what it needs for strategy execution and regime classification.
Macro & Sentiment Broader market context — fear/greed indices, dominance metrics, news sentiment — comes from public APIs and aggregated feeds. This data informs the regime classifier and strategy selection.
News Integration Crypto and traditional financial news flows through RSS feeds and news APIs, filtered for relevance before surfacing to the AI Council for commentary.
AI & Persona Layer
Large Language Models The AI Council personas are powered by large language models accessed via API. They consume structured data from the Council Data Bus and generate contextual, personality-driven responses.
Knowledge Bases Each Sage maintains a curated knowledge base of terminology, concepts, and frameworks — structured data that grounds their responses in accurate information rather than pure generation.
Discord Integration All persona interaction happens through Discord's bot framework. Messages, embeds, threads, and reactions are handled through official Discord APIs.
Communication Layer
Council Data Bus The central nervous system — a message bus architecture that carries events between trading bots, intelligence modules, and personas. Components write events; other components subscribe to what they need.
Database Backbone Shared state and historical records live in databases accessible to all components. This ensures consistency and provides the audit trail that makes the system transparent.
Platform Layer (Phase 2+)
Website The public face of Ascension — lore, graphics, documentation links, and the entry point for new members. Will host links to GitBook documentation, Discord, and the dApp once live.
dApp (Decentralized Application) Token minting, rental payments, and wallet verification will be handled through a web-based dApp. Members connect their wallets to acquire Sovereigns and manage bot rentals.
User Dashboards Authenticated dashboards (via secure login or bot-generated access links) will allow members to:
View their personal bot performance and trade history
Monitor their active rentals and configurations
Access Ascension-wide bot performance (aggregated, anonymized)
Manage their rental settings and risk parameters
Data Privacy User dashboards show your data. Members cannot view other users' individual trading data or configurations. Aggregated, anonymized performance data may be shared to demonstrate system performance, but individual privacy is protected.
Security Considerations
API Key Isolation User exchange API keys (for bot renters) are handled with appropriate security measures. Keys are stored securely, transmitted encrypted, and scoped to minimum necessary permissions.
Access Control Discord roles and token verification (via CollabLand or similar) gate access to features and channels. The system knows who should see what.
No Custody Ascension never holds user funds. Bots connect to your exchange account via API; your capital stays on the exchange under your control.
Data Privacy User trading data is private by default. Your dashboard shows your data; you cannot see other users' individual performance or configurations. Aggregated system performance may be shared publicly, but individual member data is protected.
Design Principles
Proven over novel. We use technologies with track records, active communities, and clear documentation. Debugging at 3 AM is easier when Stack Overflow has answers.
Modular over monolithic. Components are separated by function. A problem in one persona doesn't crash the trading bots. A new strategy can be added without rewriting the intelligence layer.
Observable over opaque. Logging, dashboards, and the AI Council itself exist to make the system legible. If something goes wrong, we want to know what, where, and why.
Automatable over manual. Every process that can be automated should be. The goal is a system that runs itself — not one that requires constant human intervention.
Summary
Infrastructure
Cloud hosting, process management, containerization
Trading
Bot execution, exchange connectivity, trade logging
Intelligence
Market data, macro signals, news feeds, regime classification
AI & Personas
LLM-powered Council, knowledge bases, Discord integration
Communication
Council Data Bus, shared database state
Platform
Website, dApp, user dashboards, token management
Security
API isolation, access control, non-custodial design, data privacy
The stack is designed to be boring in the best way — reliable, maintainable, and built to run without anyone watching.
Last updated