Jiwari and Kebiki
A decision reads case data from Jiwari through its API, so every caller evaluates the same facts instead of assembling their own.
How it fits
Where Archiveris sits in your architecture.
Jiwari, Kebiki and Teban hold the enterprise state, the policy determinations and the conditions on action that today live scattered across application code. Archiveris composes them into governed Business Operations, which the systems you already run invoke. Each product is still adopted on its own.
Business capabilitiesEnd-to-end outcomes
Claims processingProcurementHuman resourcesContract management
implemented and orchestrated by
Process orchestrationOutside Archiveris
Workflow toolsBPM suitesAI agentsYour applications
which call governed operations over REST, MCP and SDKs
Governed Business OperationsArchiveris composes
Validate claimAssess claimDetermine settlementApprove claim
composed from
Core primitivesWorld, logic, affordances
Jiwari
Kebiki
Teban
authorised actions execute in
Your systemsWhere the work lands
ERPCRMPolicy administrationPayments
Separately
What each product needs, and what it gives back.
Each product answers one question, takes its own kind of input, and returns an answer your systems can act on. None of them executes your business transactions: they tell your systems what is true, and what may be done.
| Product | The question | What it needs | What it returns |
|---|---|---|---|
| Jiwari | What exists, and how is it connected? | A schema for the domain, and data connected to it, typed in or imported from your systems | Governed records and relationships, through a REST API and MCP, under one permission model |
| Kebiki | What is true under policy? | The policy document, and the case data for each call | A determination with reason codes, computed facts and a trace back to the source clause |
| Teban | What may happen next? | The policy document, and the actor, resource and state for each call | The available actions with reasons, limits and the evidence that would change the answer |
Together
Each one makes the next more useful.
Jiwari holds the state that a decision or an evaluation is about. Kebiki turns policy into determinations over that state. Teban answers what may be done, and delegates the determinations it depends on to Kebiki, recording the call in its trace.
Used together they close a loop: what exists, what is true about it, and what can happen next. Used separately, each still answers its own question.
Kebiki and Teban
Teban delegates a determination, such as which approval route applies, to a Kebiki package. If it is unavailable, the action becomes awaiting review rather than allowed.
Jiwari and Teban
The context Teban evaluates describes the state of a resource. Jiwari is one convenient source of that state, though the caller may supply it from any system.
All three, with your agents
An agent can read governed context, ask what is true, and ask what it may do next, through MCP, without holding more authority than the person it acts for.
Dependencies
No product requires another.
A common worry about a portfolio is that it is really one large commitment. It is not.
- Jiwari runs without Kebiki or Teban, and holds data for whatever applications you build.
- Kebiki takes case data from any system that can send it.
- Teban evaluates context from any system, and only calls Kebiki where a policy delegates a determination.
Your stack stays yours
Keep your own interfaces, frameworks and hosting. Kebiki and Teban are called over REST or MCP. Jiwari can run against a database you operate, so records stay inside your infrastructure.
Start small, stay small if you like
One schema, one policy or one action space is a complete first step. Adding a second is a decision you make later, on evidence.
Deployment and integration vary by product and by environment. Tell us what you run and where your compliance boundary sits, and we will confirm exactly what applies to you rather than pointing at a matrix.
Adoption
Start where the foundation is weakest.
Three starting points, each worth doing on its own. Our team can take the first one from a scoped problem to production with you.
A new application that needs shared data
Model the domain in Jiwari, connect the data, and let your developers and coding agents build against one governed model with an API and MCP surface.
A policy decision that needs consistency
Compile one policy into a Kebiki package, run it alongside your existing logic, and compare the answers before you cut over.
An automated action that needs limits
Put the conditions on one action into a Teban package, and have your application or agent ask before it acts.
Bring your architecture. We will show you where it fits.
Tell us what you are building or automating, and what you run today. We will map the pieces onto your systems, honestly, including the parts you do not need.