One version
Every customer runs the same codebase. No long-lived customer branches, so upgrades always stay open.
Deployment & Pricing
The whole agent platform, deployed in your environment
KybWork is on-premises only: agents, tools, data and inference all stay inside your network boundary. Quotes depend on user count, connected systems and agent scope, with every line item spelled out.
Architecture
From model to tools, the entire agent chain stays inside your network
Delivered as containers, with every component in your data center or designated private cloud. Agents reach business tools only through the platform, and nothing calls external services at runtime.
Your network boundary
No outbound dependencies
All your business systems
- HRM
- OA
- PMS
- CRM
- ERP
- Access
- Browser
- Mobile app
- Gateway
- HTTPS gateway
- Rate limiting
- KybWork platform
- Agent runtime
- Tools & connectors
- Business services
- Data & audit
- Business database
- Vector database
- Object storage
- Audit logs
Your model service
- Public LLMs
- Private LLMs / SLMs
Container delivery
Pre-built images with a standardized deploy and upgrade process.
Starts with one database
The base setup needs a single database. Add a cache only when you need multi-replica high availability.
Swappable models
Works with any private inference service exposing an OpenAI-compatible API. No vendor lock-in.
Runs offline
License checks, update checks and telemetry are never required to run.
Pricing
Four components, no hidden project fees
- 01
Platform license
Tiered by users or nodes
Perpetual or annual subscription, covering the platform and the business workspaces you purchase.
- 02
Implementation & integration
Priced per standard component
Itemized by the number of standard connectors and deliverables, never bundled as a vague "implementation service".
- 03
Annual maintenance & upgrades
Yearly
Ongoing releases, bug fixes and technical support. Every customer runs the same version.
- 04
Add-on packs
As needed
Connector packs, industry agent packs and more, delivered as standard components at any time.
You procure hardware and compute
Servers and GPUs are never bundled into our quote. Use your existing procurement channels; we provide a recommended configuration list and connect to your existing LLM service.
Principles
A product company, not a project shop
Configuration first
Fields, approval flows, roles, dashboard metrics and connector mappings are all handled through configuration.
Connectors are product
New connectors go onto the product roadmap and are shared by every customer, not built as one-off deliverables.
FAQ
Before you deploy, you might want to know
Does it need internet access after deployment?
No. KybWork runs fully in environments with no internet egress. License checks, update checks and telemetry do not depend on external networks.
Which LLMs are supported?
KybWork connects to inference services that expose an OpenAI-compatible API. You can use a model already deployed on your network and switch models later without changing business configuration.
Can an agent exceed its permissions or make a harmful change?
Agents call tools as the requesting user and can only use allow-listed, fixed business actions. Money and HR operations require human confirmation or an approval chain, and every call is written to the audit log.
Do we need to replace our existing systems?
No. KybWork reads data from your existing systems through standard connectors and adds Q&A and collaboration on top. There is no need to migrate or replace your ERP, CRM or other systems.
Can we start with a single business domain?
Yes, and we recommend it. Pick one domain, connect three data sources, agree on one measurable metric, and deliver a first verifiable result on an 8-week plan before expanding.
What hardware do we need?
It depends on user count, the number of connected systems and how your model service is deployed. After we understand your needs, we provide a recommended configuration for your scale.
Is there a mobile app?
Yes. The mobile app supports location check-in, approvals, tasks, schedule and messages, connected to the same on-premises deployment.
Book a demo
Get your first agent running inside your network
Tell us which systems you run and what work you want to hand to agents first. We will prepare a demo and deployment plan for your environment.
