Documents survive. The reasons behind decisions don’t.
Kroma Key connects decisions, documents, people, and AI work in one knowledge map. Even when ownership changes, the reasoning stays with the organization—so people and AI don’t have to solve the same problem twice.
- Connect documents, meetings, owners, and AI work in one map
- Connect the reasons behind decisions
- See who changed what, when, and why
- Track changes
- Choose based on data location and security requirements
- From cloud to on-premises
It is a knowledge map that connects your company's knowledge
Like a transit map connecting stations, an ontology connects decisions, documents, meetings, owners, and AI work in one map. The reasons behind decisions remain with the company when ownership changes. Individuals can also connect scattered notes and AI conversations so they can find them later.
- Why did I choose this laptop model last year?
- Who approved this document, and in which meeting?
- Is that decision still valid today?
- Who is it for?
- Business teams (B2B) and individuals (B2C)
- What does it retain?
- Decision rationale, related documents, owners, and change history
- How do I start?
- Use the cloud or install it on company servers
- What improves?
- Handoffs, audit readiness, and the context available to AI
Open a decision to see everything connected to it
Each node in the graph represents a decision, document, meeting, AI analysis, or change record. Open a major decision such as the pricing policy update to see its notes, meeting records, AI analysis, and change history together.
Ask the knowledge map
Ask about a team decision
Ask in natural language, and Kroma Key finds the supporting evidence in the map. Start with one of the examples below.
Records, search, and AI automation on one graph
Records, search, audit support, and AI automation work together on the same graph instead of in separate tools.
Decision graph
Each decision becomes a node connected to documents, meetings, owners, and evidence.
Change history
Record what changed and why for handoffs and audit support.
Ontology Bot
AI organizes tags, gathers context, and proposes corrections for inconsistent records.
Context-aware AI routing
Route a question to OpenAI, Anthropic, or a local LLM. Keep sensitive work inside your environment and use external models where appropriate.
Use the ontology from chat
Connect it to messaging environments such as Slack, Discord, Telegram, and KakaoTalk. Capture decisions and their rationale where the team already works, then retrieve past decisions with supporting evidence.
- Organize tags — Identify topics and owners in conversations and organize them with tags.
- Gather context — Identify decisions and rationale worth retaining and organize them in the ontology.
- Correct records — Find inconsistent or missing records and propose corrections.
Why did we choose this pricing policy?
The team considered margin and competitor positioning. Here are the related meeting and supporting documents.
Sources: pricing policy decision · February monthly meetingAvailability, permissions, and retention policies vary by deployment. You also choose what the bot can read and retain.
Add a note to Inbox. Get the relevant context.
Add a note or task to Inbox, and Kroma Key gathers related materials and context within the scope you set, then organizes them in the ontology. You don’t have to repeat the same search every time.
Add it to Inbox
Capture a note or task as it comes to you.
Research begins
Kroma Key gathers related material and context within the scope you set.
It joins the knowledge map
The results are linked in the ontology so you can find them later.
- Starting an online store
- Notes from a tax-advisor call
- Initial marketing ideas
- Competitor research
We gathered material related to this note within the scope you set.
- Marketplace vs. storefront fee comparison
- Business registration guide
- Initial inventory and cash-flow tips
You control what Kroma Key researches and how far it can look based on your deployment and permissions.
Give every new AI session your team’s context
Conversations end, but the reasons behind decisions remain in the ontology. Use the Ontology API to pass that context to AI tools such as Claude Code and Codex, so a new session can answer using your team’s terminology and decisions.
- Re-explain the background every time you start a new AI session.
- The reasons behind past decisions disappear when the conversation ends.
- Each person explains the context differently, so AI answers vary.
- Decisions and their rationale remain in the ontology.
- Use the Ontology API to pass only the context a session needs.
- AI answers from the same context, regardless of who starts the session.
Retain the context
Keep decisions, rationale, and terminology in the ontology.
Pass it through the API
Use the Ontology API to select and pass the context the AI session needs now.
Continue with shared context
A new session answers using your team’s terminology and decisions.
New Codex session: "Why did we choose tiered pricing?"When the Ontology API supplies the decision history, the AI can answer with evidence: "We moved from an across-the-board increase to tiered pricing to reduce churn among core customers."
# 1) Download only the context you need from the ontology
curl -s "$ONTOLOGY_CONTEXT_URL" \
-H "Authorization: Bearer $KROMA_KEY" > context.md
# 2) Start a new AI session with that context
"$AI_AGENT_CLI" --context-file context.md
# Actual endpoints and AI tool flags vary by deployment
Integration and security policies are configured for your deployment. You also choose the scope and permissions of the context you pass.
Keep data and records under customer control
Review these controls before connecting decisions and documents. They matter most to security-sensitive organizations.
Keep data under customer control
Keep data on company servers with on-premises or hybrid deployment, or control individual access through a virtual organizational structure.
Maintain an audit trail
Record who accessed or changed information, when, and why.
Validate with a PoC first
Define the scope and security review criteria before starting. You do not have to connect everything at once.
Availability and security policies vary by deployment.
The same model works for teams and individuals
For businesses
Connect team documents, decisions, meetings, and owners. Designed for organizations with frequent handoffs and audit requirements.
For individuals
Manage notes, decisions, and AI conversations in a personal knowledge map. It is cloud-based, with nothing to install.
For security-sensitive organizations
If data cannot leave the organization, keep it inside with on-premises deployment (installed on company servers) or a hybrid setup. Cloud is sufficient for many organizations, and on-premises is also supported.
When security requires it, design the hardware too
Not every customer needs on-premises deployment. For organizations where data location matters—such as finance, manufacturing, and R&D—Kroma Key can design storage, local AI, and backup policies together.
Kroma NAS
Store high-volume documents and decision history
- Laptop CPU
- DDR4 64GB Memory
- 8TB NVMe Cache
- Up to 128TB HDD
Kroma RAID Pro
Security-sensitive work and fast local AI inference
- Desktop CPU
- 128GB DDR5 RAM
- 16TB NVMe SSD
- RTX 8000 48GB GDDR6
Pay for what you need
Exact pricing depends on user count, data volume, and security requirements. The model has three tiers.
Personal Ontology
Monthly subscription
Manage notes, decisions, and AI conversations in a private cloud graph.
- Personal knowledge graph
- AI-assisted tagging
- Search and review
- Cloud-hosted
Business Workspace
Based on seats and usage
Manage team documents, decisions, and owner context together.
- Team permission management
- Ontology API
- Ontology Bot
- Context-aware AI routing
Enterprise / On-Premise
Implementation fee + monthly operations fee
Deploy inside the company with dedicated hardware and security policies.
- On-premises · Hybrid
- Service-level agreement (SLA)
- Audit trail
- Dedicated hardware
- For a proof of concept (PoC), we define the scope and security review criteria before providing a quote.
- AI model costs are calculated separately based on OpenAI, Anthropic, and local LLM usage.
- Choose to purchase hardware, lease it, or use equipment you already own.
Keep the tools you already use
Start by connecting existing documents, decisions, and meeting records in a graph that both people and AI can read. Choose the deployment model that fits your data location and security requirements.
- Individuals can start in the cloud.
- Teams can start by connecting documents and decisions in Business Workspace.
- Security-sensitive organizations can deploy on-premises or in a hybrid environment.
- Use the Ontology API and Ontology Bot for supported repeatable workflows.
Start with the questions that matter
Keep your current workflow. Start retaining the reasons behind decisions.
Tell us your team size and data environment, and we will recommend a deployment model. A PoC begins after we define its scope and security review criteria together.