Customers
Customer AI
Assistants & intake
Answers, qualification, structured leads
Open-source AI engineering studio
Practical AI systems, built on open source.
We build focused AI tools for businesses — customer assistants, private team knowledge systems and workflow automation — using our own open-source infrastructure.
- Customer assistants, private team knowledge, and workflow automation — built for a specific business.
- Client work starts on open-source infrastructure we maintain ourselves.
- We don't sell a generic AI platform. We build the useful part for your business.
Studio map
One workflow entering a custom build
Routing from your process
Stage 1 · Your workflow
Questions, documents, or a process that repeats
Stage 2 · Open-source foundation
Reuse the tools we already maintain
Stage 3 · Custom build
Write only the part specific to you
Configured around your data, tools, and handoff
Team
Team Knowledge
Private workspace
Docs, onboarding, cited answers
Ops
Workflow Automation
Repeating processes
Handoffs across the tools you already use
Foundation
Open Source
Infrastructure we maintain
Knowledge, agents, observability, UI
Development packages
Three kinds of system. Built for your business.
These are typical project classes — not SaaS products. We reuse the same open-source foundation and write the part that is specific to you.
Customers Customer AI
Team Team Knowledge
Process Workflow Automation
01
Customer AI
Customer-facing assistants that answer questions, understand enquiries and collect structured leads.
- Visitor asks a question or describes a need
- AI answers using company knowledge
- Asks clarification when necessary
- Collects structured information
- Qualifies the enquiry
02
Team Knowledge & Onboarding
Private AI knowledge systems that help employees find answers and learn how your company works.
- Company documents, policies, guides, SOPs and FAQs
- Structured searchable knowledge layer
- Private AI assistant
- Employees ask questions
- Answers grounded in company sources
03
AI Workflow Automation
Focused automation for repetitive workflows involving messages, documents and information.
- Show us one repetitive workflow
- We identify what can be automated
- We reuse proven components
- We build the missing steps and integrations
- We deploy and iterate from real usage
How we work
Every implementation is different. The infrastructure doesn't have to be.
Every system is configured and extended around your data, workflows and existing tools.
Delivery flow
From the job to be done to a system in your environment
Reuse first, then write the gap
- 01 →
Understand the workflow
We start from the job to be done — the questions customers ask, the documents employees need, or the process that repeats.
- 02 →
Reuse proven open-source components
Client projects start from the same tools we maintain for ourselves: knowledge, agents, observability and admin.
- 03 →
Build the missing UI and integrations
We provide the foundation and build the business-specific layer around it — data, workflow and existing tools.
- 04 →
Deploy and test in your environment
Self-host or deploy with us. We are not locked to a single AI provider, and neither are you.
- 05
Iterate from real usage
The first version is meant to be used. We tighten prompts, retrieval and handoff from actual traffic — not from a slide deck.
Built on open source
Independent tools we use in client work
Our client work starts with the same open-source tools we build for ourselves. That means less time rebuilding infrastructure and more time solving the part specific to your business.
Discuss a project
Tell us what you are trying to build. We will look at the workflow and say what can reuse existing infrastructure — and what we would write for you.
Start a conversation
A rough description is enough
Name the workflow if you already know it. No generic platform trial — this is a project conversation.
Discuss a project
A rough description is enough.