Run AI agents across your enterprise, on your own terms
Kog Weave is a complete, modular platform for building, running and governing AI agents. Deploy it in your own cloud or data centre and connect it to the systems your business already runs on.
Deploy on
Where people work
- Microsoft Teams
- Slack
- Chrome
- Your own apps
Kog WeaveIn your cloudWhat agents work with
- CRM & ERP
- Email & calendars
- Documents & knowledge
- Cloud AI models
- Private models
Why Kog Weave
Everything an enterprise needs to run AI agents, in one platform
Most companies stitch together models, agent frameworks, connectors and security tools themselves. Weave brings them into one governed platform that your IT team can run.
Runs in your environment
Install Weave in your own cloud account or data centre. Your data, prompts and conversations stay on infrastructure you control.
Pluggable by design
Every part of the platform is a building block. Swap models, agents and tools without rebuilding what you already have.
Governed from day one
Built-in guards for personal data and prompt injection, separation between teams, single sign-on and a full trail of what every agent did.
Built to keep costs down
Smart model routing and token-lean tools mean you pay for the intelligence you need, not for waste.
The platform
Building blocks, woven together
Use them as one platform or pick the pieces you need. Each block is independent and replaceable, so Weave fits the way your organisation already works.
Agent Gateway
The front door and control room. Runs every agent for every team or customer, safely separated, under one consistent set of rules.
Pluggable Agents
Choose the agent that fits the job, or bring your own. Several agent engines ship in the box, and each one can be replaced.
Context & Memory
Agents remember what matters. Your documents become searchable knowledge bases, and skills teach agents how your company works.
Tool Gateway
Connect agents to your business systems through one secure gateway. It finds the right tools on demand and checks every call and every answer.
Governed Inference
One entry point to every AI model. Send each request to the best-fit model, switch providers with a setting, and screen what goes in and out.
Private Models
Run open models on your own hardware, on virtual machines, Kubernetes or OpenShift, for work that must never leave the building.
Ready-made Connectors
Built-in connectors for the major business systems you already use, tuned to use as few tokens as possible.
Identity & Access
Sign in with your corporate identity provider. Access to connected systems is kept in a secure vault, never handed to the model.
Agent Observability
Trace every step an agent takes, measure quality over time and find problems before your users do.
Admin Console
Create and configure agents, try them in a playground, run evaluations, follow usage and cost, and review logs in one place.
Skills & Knowledge
Upload policies, manuals and playbooks. Weave turns them into knowledge that agents search and reason over, and into skills that teach agents to follow your way of working.
Channels
Agents that meet people where they already work
One agent, many front doors. Configure an agent once and make it available in the tools your employees and customers use every day.
Microsoft Teams
Agents in the chats and channels your people already use.
Slack
Ask, delegate and get updates without leaving Slack.
Serve customers and field teams on the go.
Chrome extension
An assistant beside every web app, aware of the page you are on.
Embeddable widget
Add an AI assistant to your own product with a few lines of code.
Start where you are
Adopt the whole platform, or just the part you need
Every Weave service can be deployed on its own. Start small, prove the value, and grow into the rest.
The full platform
Gateway, agents, memory, tools, models and the admin console, working together as one enterprise AI agent platform.
Just the Tool Gateway
Already using a hosted assistant such as ChatGPT Enterprise? Give it governed access to your systems, with connectors, sign-in and guards included.
Just Governed Inference
Building your own agents? Point them at one governed model endpoint with routing, guards and usage tracking built in.
Deploy anywhere
Your cloud. Your data centre. Your rules.
Weave runs wherever your governed infrastructure lives, from a public cloud account to a fully private OpenShift cluster, including private AI models on your own hardware.
Install the way your IT team prefers
Coming soonRed Hat OpenShift Operator
Install and upgrade Weave the OpenShift way.
Azure Marketplace
Deploy into your Azure subscription and buy through your existing agreement.
AWS Marketplace
Deploy into your AWS account with consolidated billing.
Security & governance
Guardrails built in, not bolted on
Agents only become useful when they can reach real systems and real data. Weave makes that safe, with protection on every request, every tool and every answer, and a clear record your compliance team can rely on.
Talk to us about security- Personal data is detected and masked before it reaches a model
- Prompt injection is blocked on the way in and on the way out
- Every tool call is checked before it runs, and every result before it is used
- Teams and customers are kept apart from each other
- Single sign-on with your enterprise identity provider
- A complete, searchable trail of every agent decision
- Your data never has to leave your infrastructure.
See Kog Weave running in your environment
We will walk you through the platform, map it to your infrastructure and help you pick the first agents worth building.