Pricing built around your business, not your bill.

Yes. We built the catalogue before opening for clients. Every automation is scoped, designed, and engineered on Anthropic Claude and self-hosted n8n. Some run as live demos on our own infrastructure, others are deployed to client environments. Your setup configures the ones you pick onto your real data and connects them to your systems.

A one-off setup fee plus a monthly retainer. Starter is £250 setup and £199 a month for up to 5 automations; Growth, Pro and Custom scale up from there. The retainer covers hosting, monitoring, Claude processing within a fair-use band, business-hours support, and ongoing tuning. Full breakdown on the pricing page.

No lock-in. The retainer is billed monthly with 30 days notice to cancel. A 6-month commitment gets you 1 month free — roughly 15% off the year-one total — but it is optional. You own the workflows and configuration either way.

Our first 5 clients get 50% off setup and the first 2 months free, in exchange for a written case study at month 3. It is how we build proof while you get the work at a discount. Mention it when you book a discovery call.

That is what the free 20-minute discovery call is for. We map the work, recommend the tier that fits, and show you the saving — the hours of admin you get back — before anything is signed. We do not run generic public demos because each automation has to be wired into your specific data and systems to mean anything.

Most of the engineering is already done, so deployment is measured in days, not quarters. After the discovery call we configure your chosen automations onto your data, harden them for production, and go live. Larger Custom rollouts are scoped in days and quoted upfront.

If you have repetitive work that follows rules, you are ready. The free discovery call removes the guesswork — we look at your actual workflows, tell you honestly which ones are worth automating, and recommend a tier. No pressure to proceed.

Real examples from our catalogue: drafts replies to common customer enquiries grounded in your knowledge base; chases unpaid invoices on a schedule and escalates when needed; generates a weekly business summary in plain English from your CRM, accounts, and inbox; books appointments from inbound emails without a human reading them. Each automation is scoped to one job, done well.

Your retainer absorbs reasonable API cost fluctuations within its fair-use band. If there is a structural change to underlying model pricing, we are transparent about it and re-quote with notice. We also use Haiku for simple work and Sonnet only where reasoning depth matters, keeping costs predictable.

On self-hosted n8n on our managed infrastructure by default — we host the workflows, you connect your systems. If you need full data residency on your own servers, we deploy n8n into your infrastructure instead. Either way, no client data is sent to model training, and everything stays in the UK unless you ask otherwise.

On your infrastructure if you self-host, or on Zyntho infrastructure in UK data centres. Either way: no third-party training on your data, no transfers outside the UK without your explicit consent, full GDPR compliance, and a Data Processing Agreement signed before any data is touched.

Every automation is built on open standards — n8n (open source) and the Anthropic API. You own the workflows and the configuration. We document everything so any technical operator can take over. The retainer is the cost of us running it — but you can run it yourself, hire someone else to, or pause it any time. No vendor lock-in.

Cancel the retainer with 30 days notice. We provide all configuration, documentation, and a handover call. The automation either continues on your infrastructure or is decommissioned cleanly. Your data stays with you. This is in writing in every contract.

You own the configuration, the prompts written for your business, the workflow logic, and all outputs. We retain the underlying patterns and components that are reused across clients. This is industry-standard and documented in the Master Services Agreement.

n8n is open source, self-hostable, and significantly cheaper at scale than Zapier or Make. It gives us deeper control over data flows, custom logic, and reliability. We can host it on your servers if you require full data residency. Zapier and Make are competent for simple connections but constrain what serious AI workflows need.

For long-context reasoning, tool use, and the kind of careful instruction following business automation needs, Claude has consistently been the model we trust most. Anthropic's safety and governance posture matches how we think about deploying AI responsibly. We can integrate other models where they fit better — but Claude is our default.

Yes. We are a UK-based business operating under UK GDPR. Every engagement starts with a Data Processing Agreement. No client data is used to train models. Data residency stays in the UK by default. We have not yet pursued formal accreditations like Cyber Essentials Plus or ISO 42001 — those come as the business scales. The substance of compliance is in place from day one.

Yes, before any data is shared. Mutual NDA template available on request, or we sign yours.

We integrate, extend, or replace based on what makes sense. If you have working Zapier flows, we can connect to them. If you have legacy systems, we bridge them. The discovery call is where we work out the right move — we will not recommend ripping anything out unless the case is obvious.