Tools we built because they were missing.
Every tool on this page came out of a real project, not a product plan. Somewhere a piece of software was missing, the market had no usable answer, so we wrote it ourselves.
Why does a studio build its own tools?
Anyone building AI systems for clinics and SMEs, as RakenAI does, quickly notices where their own workflows jam. You can bridge those spots with manual work. Or you build the missing piece, run it yourself, and know first-hand what it costs to keep a system like that alive in daily use.
We would rather build. It keeps us close to what is changing in the field: models, protocols, agent standards. And it has a side effect that matters to clients. Every recommendation we make, we have tried on ourselves first.
Six of them are below. Three are publicly reachable and linked. Three run internally or are being built; we describe those without pretending you could click into them today.
Six tools, three of them public
RakenAI Audit
A visibility and demand audit for owners: domain in, a plain-language brief out. Built because classic SEO reports leave operators alone with raw data, and the question of what any of it means for revenue stays open.
The tool is being rebuilt and is temporarily unreachable. Our audit page still takes requests; access follows as soon as it is ready.
Agent Tender
A precision radar for public tenders. Public procurement is scattered across more than 60 platforms in more than 30 languages, and the established tools cover the US and the EU while barely reaching Southeast Asia, the Middle East and Central Asia.
Agent Tender reads the original sources, qualifies every hit against the supplier's own profile and delivers it by mail with traceable exclusion reasons; checking, pricing and submitting stays with the company. Currently in closed operation.
Chatbot for Clinics
A chat system for clinics and SMEs in the DACH region, currently in build. Channels: WhatsApp, Instagram, Facebook and a web widget. At its core sits a separate knowledge base per client: the bot answers from that clinic's own data rather than from generic model knowledge.
The first pilot is HLC Hairline Clinic, whose system work is described on our references page.
Citegate
Researched, sourced, compliance-checked web and blog content in the brand's own voice, built for clinics, law firms and regulated industries.
It grew out of a problem we saw daily: generic AI invents services, sources and promises that a regulated business is not allowed to make.
webmcphelp
Makes websites usable by AI agents through the WebMCP standard, after checking whether the site can carry it at all.
Built when it became clear that agents will book, inquire and compare, and most websites give them no interface for it.
Second Brain Kit
An open Obsidian template for working with AI agents, MIT-licensed on GitHub: a map of the knowledge base, an operating manual for agents, a handoff protocol per project.
The template ships empty on purpose, with no private data and no invented examples. It is the pattern we work by ourselves.