Problem99
Problem99 builds orchestration infrastructure for AI agents and agentic workflows. Its stated vision is a human-led, AI-executed way of working, with unified environments that connect intent to action rather than replacing people.
At a Glance
- Solo software developers
- Software-development teams and product teams
- Organizations adopting AI coding agents and agentic workflows
- Enterprise engineering organizations needing custom deployment, SSO, security, and support
- +1 more
AI Tools by Problem99
(1)
kodeblue
AI Blue Team Agent for Linux
Discussions
No discussions yet
Be the first to start a discussion about Problem99
Latest News
Kodeblue fixed false self-integrity alerts during hub-started upgrades.
Kodeblue added operator-controlled server updates from the hub and improved containment-rule validation and approval safety.
Kodeblue gave vulnerabilities and auto-block rules dedicated server tabs in the hub.
Kodeblue improved one-command release installation, hub installation, model setup, and fresh-install behavior.
Products & Services
An integrated agentic software-development workspace combining team boards, planning, timeline and messenger, a browser IDE, coding agents, isolated Git worktrees, previews, automation, cloud development environments, model access, analytics, and code review. Each task gets a Taskie that can coordinate specialist Subies and share project knowledge.
A centralized MCP-compliant API and long-term memory for AI agents. It provides a verified solution database for coding errors, community solution submission and voting, and token-optimized vendor-documentation retrieval; it integrates with Claude Desktop, VS Code, Cursor, and custom agents.
An Apache-2.0 open-source AI blue-team analyst for Linux servers. It learns a host's normal state, watches for changes, triages and investigates incidents, and can contain attacks under policy; an optional self-hosted hub manages fleets, incidents, approvals, reasoning, and Slack delivery.
Market Position
Problem99 positions Kodeboard as an integrated alternative to fragmented planning, coding-agent, IDE, preview, collaboration, and automation stacks: one subscription combines the workspace and AI credits, while Taskies/Subies preserve shared project context. Kodeblue differentiates from conventional alert-only monitoring and human-only SOC workflows by combining deterministic sensors with per-server AI investigation and policy-controlled containment. Hopper positions itself as verified, token-efficient agent memory rather than open-web searching.
Leadership
Founders
Ram Margulies
AI expert and entrepreneur; CEO and founder of Problem99 and CTO of SijakMedia. LinkedIn search results also describe a background as a SecOps specialist with infrastructure penetration-testing and detection-engineering experience.
Mihai Neacsu
CRO and co-founder of Problem99; also co-founder/CRO at SijakMedia and previously a business-development consultant at Adscope, according to LinkedIn search results.
Executive Team
Ram Margulies
CEO & Founder
AI expert and entrepreneur; CTO at SijakMedia and a SecOps specialist with infrastructure penetration-testing and detection-engineering experience.
Mihai Neacsu
CRO & Co-Founder
Business-development executive; co-founder/CRO at SijakMedia and previously a business-development consultant at Adscope.
Founding Story
Problem99 says it is building a new way of working in which humans lead and AI executes. Kodeboard arose from the team's experience with fragmented development tools and repeated agent exploration: they wanted planning, coding agents, shared project knowledge, and review in one workspace so agents could carry context forward and collaborate around a project.
Business Model
Revenue Model
Kodeboard uses paid monthly or annual subscriptions that bundle a cloud workspace, seats, and a monthly AI-credit budget; extra credits are available. Enterprise uses custom credit volume, deployment, security, and support. Hopper uses API credits for paid search/upload and documentation queries, while Kodeblue is open-source and self-hosted.
Pricing Tiers
One seat; $25 in monthly AI credits; up to 5 concurrent tasks and 2 concurrent projects; 4 GB RAM and at least 2 shared CPU cores.
Two seats and $100 in shared monthly AI credits; 2-25 seats with pooled usage; up to 10 concurrent tasks per seat; 8 GB base RAM scaling to 12 GB.
Custom credit volume, unlimited seats, custom deployment, SSO and advanced security, SLA and dedicated support, and a dedicated account manager.
Target Markets
- Solo software developers
- Software-development teams and product teams
- Organizations adopting AI coding agents and agentic workflows
- Enterprise engineering organizations needing custom deployment, SSO, security, and support
- Small and mid-sized teams needing Linux-server security monitoring without a full 24/7 SOC
- Teams planning, building, reviewing, and shipping software with AI coding agents
- Solo developers using a managed cloud workspace with agent-assisted projects
- Multi-person engineering teams sharing AI credits, tasks, projects, and collaboration context
- AI agents querying verified fixes for coding errors through REST or MCP
- Linux-server SOC monitoring, incident response, vulnerability assessment, and SecOps for teams without a 24/7 SOC