Ejentum
Ejentum provides a reasoning harness for agentic AI to prevent cognitive drift and improve task reliability through structured reasoning injections.
At a Glance
- AI Agent Developers
- Enterprise AI Safety Teams
- Open-source AI Builders
- Competitive Programming Platform users
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
Products & Services
A pre-generation cognitive scaffold for AI agents that provides reasoning, code, anti-deception, and memory harnesses.
An implementation of the Model Context Protocol (MCP) providing 679 cognitive abilities to tools like Claude and Zed.
Market Position
Positions itself as the 'last mile of agent reliability', moving agents from high 80s to near 100% success rates by injecting discipline rather than more intelligence.
Leadership
Founders
Frank Brsrk (Franko Luci)
Self-taught AI developer and solo founder.
Executive Team
Frank Brsrk
Founder & Engineer
Self-taught developer responsible for engineering 679 abilities.
Robert Franks
Lead Engineer (Applied AI Persona)
Research funder and technical lead, described as an Applied AI Persona.
Board of Directors
Founding Story
Founded in March 2026 to address the need for discipline and reliability in AI agents. The founder built the system to ensure agents don't hallucinate or drift from their tasks.
Business Model
Revenue Model
Subscription-based (SaaS) for API usage and adaptive calls.
Pricing Tiers
1,000 dynamic calls/month, no adaptive, 30 days.
1,000 dynamic + 250 adaptive calls/month.
5,000 dynamic + 1,500 adaptive calls/month.
Target Markets
- AI Agent Developers
- Enterprise AI Safety Teams
- Open-source AI Builders
- Competitive Programming Platform users
- Competitive programming
- Complex legal research
- Multi-step decision making
- Adversarial auditing of AI agents
- Preventing sycophancy and social engineering in LLMs
- Individual developers using n8n and MCP-compatible IDEs like Zed.