The chatbot framework for enterprise AI

AI assistants that live inside your cloud.

Ollie is a licensable chatbot framework that brings memory, context and graph intelligence to AI — the foundation for customer service bots, agentic systems, and personalisation. Deployed in your own AWS, Azure or GCP, across web, app and voice.

Runs in your cloud Memory · Context · Graph Web · App · Voice License + build services
Built for teams in
Financial Services Telecoms Healthcare Retail Public Sector Travel Insurance
What you get

A framework you license.
A team that builds with you.

Ollie isn't another SaaS subscription you'll outgrow. It's the foundation your AI runs on — owned, hosted and operated inside your business — with our team alongside yours to ship the first solution and scale the next.

License the framework

Pay once for the foundation. Memory, context, graph, tools, guardrails and channels — yours to run forever, with version updates included.

We build the solution

Our engineers partner with yours to design, build and launch the bot, copilot or agent — fixed-scope projects, not open-ended retainers.

Your cloud, your data

Ollie deploys inside your AWS, Azure, GCP or on-prem environment. Your data, your model keys, your audit trail — never ours.

Platform

Memory, context and graph —
the three things AI is missing.

Most chatbots forget who you are, miss what just happened, and can't reason across your business. Ollie is the framework that fixes all three, and bundles the rest of what production AI actually needs.

Memory

Durable, per-user memory that survives sessions, channels and deployments — so customers stop introducing themselves on every call.

Context

Retrieval over your documents, databases and live events. Answers grounded in your truth — not the model's training data.

Graph

Model the relationships between customers, products, accounts and events. The bot reasons over your domain, not just over text.

Tools & actions

Connect to CRMs, ticketing, billing and internal APIs. The bot doesn't just answer — it transacts, books and resolves.

Guardrails & policy

PII redaction, jailbreak defence, per-tenant limits and policy routing — enforced at runtime, not improvised in prompts.

Observability

Conversations, tool traces, costs and quality metrics — in your stack. See what the bot did, why, and what it cost you.

Solutions

One foundation.
Every kind of assistant.

Ollie is the platform underneath your highest-value AI products. Customer service today, agentic workflows next quarter, personalisation across the year — the same framework, the same data, compounding.

Customer Service

Service bots that resolve

Deflect tier-1 tickets, action refunds and bookings, and hand off to humans with full context and history — across web, app and the phone line.

Agentic AI

Agents that take action

Long-running agents that research, decide and execute across your tools — with checkpoints, audit trails and human-in-the-loop approvals.

Personalisation

Personalisation engines

Use the graph and memory layer to power recommendations, next-best-actions, and per-user experiences inside your existing products.

Internal

Employee copilots

Give every team a copilot that can query data, run workflows and remember context — with policy and access control wired to your IdP.

Vertical

Domain assistants

Healthcare intake, financial advice, legal triage, telecoms support — production scaffolding so you skip the plumbing and ship judgment.

Embedded

AI inside your product

Drop a fully-themed assistant into your app or website. Branded, isolated per-user, and instrumented from day one.

Channels

Web, app, and the phone line — one brain.

Most vendors give you a chat widget. Ollie is the framework that powers every channel where your customers actually are, with a single source of memory, context and policy behind them.

  • WebEmbeddable widget or full-page experience, themed to your brand and shipped behind your auth.
  • Mobile appNative iOS and Android SDKs with offline buffering, push handoff, and on-device redaction.
  • Voice & VoIPReal-time speech, SIP/WebRTC integration, barge-in, warm transfer to a human — wired to the same memory as your chat.
  • One conversationA customer who started on the website finishes on the phone. The bot remembers — because it's the same bot.
// Drop into any web page.
<script src="https://acme.cloud/ollie.js"></script>
<ollie-assistant
  workspace="acme.support"
  user="{{ session.user.id }}"
></ollie-assistant>
100%
in your cloud
3
channels, one brain
Any
model, any provider
licence, then yours
Deployment

Hosted in your cloud.
Powered by any model.

Ollie installs into your AWS, Azure or GCP account — or on-prem if you need it. Bring your own model keys (Azure OpenAI, AWS Bedrock, Google Vertex, Anthropic, OpenAI) or use Ollie's model platform with a small per-token fee.

Your data never leaves your perimeter.

The framework deploys as containers into a VPC you own. Conversations, memory, embeddings and graph data sit in your databases. Model calls go to your provider with your keys. Compliance and audit teams stop us at the door — and we stay there.

Talk to our solutions team
Commercial model

Pay once for the foundation.
Tokens only if you use ours.

No per-seat lock-in, no surprise SaaS bills. License the framework, optionally pay us to build the first solution, and choose the model platform that suits your business.

Foundation Licence
One-time
The Ollie framework, deployed in your cloud, yours to run forever.
  • Memory, context & graph runtime
  • Web, app and voice channels
  • Tools, guardrails & observability
  • Version updates & security patches
  • Deployed into your AWS / Azure / GCP
Request pricing
Models — your choice
BYOor per-token
Run on any model provider, or use Ollie's model platform with a small token charge.
  • Bring your own Azure OpenAI keys
  • Bring your own AWS Bedrock account
  • Bring your own Google Vertex project
  • Or route through Ollie's model platform
  • Switch providers per workload, no rework
See model options
Let's build your first assistant

Own the AI layer
of your business.

Stop renting a chatbot you'll outgrow. License the framework, deploy it into your cloud, and build the assistants your customers and teams will actually use.