All solutions

Core Platform

Process management for the AI era

A platform where every operational process has a human-readable form, a named owner, full version history, and an MCP interface. Accessible to people and AI agents simultaneously.

Platform capabilities

Everything a process needs to stay alive

Markdown + YAML frontmatter

Every process is a human-readable Markdown document with structured YAML frontmatter, following the Agent Skills spec. Editable by anyone, parseable by any AI.

Immutable version history

Every save creates a new immutable version. View diffs, roll back to any point, and see which version an AI client consumed on any given call.

Named ownership

Every department and process has a named owner, a real person accountable for keeping it accurate. Ownership is the accountability anchor.

MCP delivery layer

An MCP server exposes your processes to any compatible AI client. Three tools: read_process, list_processes, and update_process for AI-assisted authoring.

Usage analytics

Every MCP call writes a structured event. See adoption rates, per-client breakdowns, staleness pressure, and skill health scores per process.

Staleness alerts

Configurable rules trigger email or webhook notifications when a process hasn't been reviewed in N days, or has been called many times without an update.

Org hierarchy

Structure that mirrors how your org works

Koinoflow organizes processes around a four-level hierarchy that maps naturally to how real companies operate. Ownership cascades down: every department and every process has a named accountable person.

1

Workspace

Acme Corp · Top-level org unit

2

Team

Engineering · Division or business unit

3

Department

Frontend · Functional area with an owner

4

Process

Deploy to production · Owned procedure, the atomic unit

Usage event log
// Every MCP call writes a usage event:
{
  "process_id": "proc_4f8a2b",
  "process_slug": "engineering/deploy-to-production",
  "version": "2.4.1",
  "client_id": "cursor-workspace-abc",
  "client_type": "cursor",
  "called_at": "2026-04-11T09:12:33Z"
}

// Aggregated analytics:
{
  "process": "Deploy to production",
  "calls_30d": 847,
  "unique_clients": 12,
  "last_updated": "2026-01-15",  // 86 days ago
  "health_score": 34,
  "alert": "staleness: review overdue"
}

Observability

Turn AI usage into actionable insight

Every MCP call writes a structured usage event. The analytics dashboard turns that raw log into adoption rates, staleness pressure, per-client breakdowns, and skill health scores.

  • "Your onboarding process was called 847 times last month"
  • "It hasn't been updated in 90 days. Here are the 3 AI clients still on v1.2."
  • "Your Deploy process has a health score of 34. It's stale and unowned."

Included in all plans.

Give your AI agents a single source of truth

Start with a free workspace. MCP server running in 30 minutes.