Ejentum
Ejentum provides an engineered reasoning harness for agentic AI, injecting cognitive abilities at inference time to prevent reasoning decay.
At a Glance
- AI Developers
- Agentic Startup Founders
- Enterprise AI Safety Researchers
AI Tools by Ejentum
(1)
Ejentum
Agent Reasoning Correction API
Discussions
No discussions yet
Be the first to start a discussion about Ejentum
Latest News
Publication of 'Under Pressure' Research Paper
Ejentum featured on Product Hunt
Integration with Official MCP Registry
Products & Services
A layer of 311 operations covering 6 cognitive dimensions including causal, temporal, and spatial reasoning.
128 operations focused on engineering discipline, correctness checks, and verification loops for coding agents.
139 operations designed to block sycophancy, hallucination, and deception in AI models.
101 operations that sharpen perception and track state/context drift during long-running tasks.
Market Position
Positions itself as a specialized orchestration and reasoning layer that adds discipline to existing LLMs rather than just increasing parameter count.
Leadership
Founders
Frank Brsrk (Franko Luci)
Self-taught AI developer and lead engineer who engineered 679 cognitive abilities for the Ejentum reasoning harness.
Executive Team
Frank Brsrk
Founder & Lead Engineer
Self-taught developer and the primary architect behind the 679 cognitive abilities.
Board of Directors
Founding Story
Founded to solve 'static cognition' issues in language models by treating failures as structural rather than informational. The project was built by a solo developer and funded by private research and infrastructure supporters.
Business Model
Revenue Model
Subscription-based and usage-based API model (Dynamic and Adaptive calls).
Pricing Tiers
30 days, 1,000 dynamic calls, no credit card required.
1,000 dynamic calls and 250 adaptive calls per month.
5,000 dynamic calls, 1,500 adaptive calls, and Hosted MCP access.
Target Markets
- AI Developers
- Agentic Startup Founders
- Enterprise AI Safety Researchers
- Reliable multi-step AI agents
- Scientific and engineering computing verification
- Anti-hallucination and security layers for LLMs
- Optimizing coding agent performance