Cosmic Stack
Cosmic Stack is a research and product lab building the agents, infrastructure, and tools of the next decade, focusing on closing the gap between technical AI possibilities and shipped products.
At a Glance
- Engineers
- Developers
- AI Startups
- Individual Finance Users
AI Tools by Cosmic Stack
(1)
Mercury Agent
Open Source AI Agent Framework
Discussions
No discussions yet
Be the first to start a discussion about Cosmic Stack
Latest News
Mercury Agent v1.1.11 released with enhanced Autopilot
Cosmic Stack Incubator officially opens for applications
Mercury Skills Registry reaches 126+ specialized skills
Sav.ink private beta launch announced
Products & Services
A terminal-native coding agent for engineers featuring multi-agent orchestration, loop detection, and a persistent 'Second Brain' memory system.
A private AI finance manager that ingests, reasons, and plans personal finances.
A managed runtime for running Cosmic Stack agents and custom agents as a service.
Market Position
Cosmic Stack focuses on the runtime layer of AI agents (tools, memory, evals) rather than just the model, emphasizing production reliability and open-source infrastructure.
Leadership
Founders
Salman Qureshi
Co-founder of Mercury Agent and lead at Cosmic Stack. Developer and open-source enthusiast known for building tools that simplify developer workflows (GitHub: hotheadhacker).
Akshat Chauhan
Co-founder and Lead Engineer/Architect at The Cosmic Stack. Specializes in AI, agents, AWS, and Databricks. Former software and product developer.
Avantika Nautiyal
Co-founder at The Cosmic Stack. Background in building scalable products, AI systems, and high-performance digital experiences (DUCS '24).
Yogesh M
Co-founder at The Cosmic Stack, focusing on intelligent software design and development for businesses and startups.
Executive Team
Salman Qureshi
Co-founder
Lead at Cosmic Stack, creator of Mercury Agent.
Akshat Chauhan
Co-founder & Lead Engineer
Architect for AI and agent infrastructure.
Founding Story
Cosmic Stack was founded because while frontier inference costs dropped significantly, AI products didn't catch up. The founders aimed to bridge the gap between technically possible and actually shipped AI agents and infrastructure.
Business Model
Revenue Model
Managed cloud services (SaaS), equity/profit-sharing from incubated products, and enterprise support for open-source primitives.
Pricing Tiers
MIT licensed, run locally.
Closed beta, managed runtime services.
Target Markets
- Engineers
- Developers
- AI Startups
- Individual Finance Users
- Autonomous coding and software engineering
- Personal financial planning and management
- Workflow automation via AI agents
- Infrastructure and tool building for developers
- Open-source community
- Incubated stealth companies