We build AI agents for small teams, and keep them running.
Agents that reply to new leads, sort the shared inbox, update the CRM and write the Monday report. We build them on Qoren or on your own stack, and look after them once they are live. Send us the job and you get a fixed price within 2 business days.
Agents on Qoren are the quickest to set up, because there are no servers to provision. When your stack or your security rules need something else, we build it there. We can maintain either kind after launch.
We set up agents on Qoren and hand them over working.
We design, configure and deploy OpenClaw, Hermes or Codex agents on Qoren, the agent platform we build and run. Each agent is built around a workflow you already have, such as replying to new leads or sorting a shared inbox, and comes with a spend cap, monitoring and a written handover.
An agent built around one workflow you already run
Hosted on Qoren, with encrypted secrets and a hard spend cap
Connected to your inbox, CRM, calendar and other tools
From $2,000
Per workflow, fixed before we start. Qoren hosting from $39 a month, billed by Qoren.
Agents built on your own cloud or the stack you already use.
When the work has to run in your own cloud, inside your data boundaries or on a stack other than Qoren, we build it there. That can be custom code in Python or TypeScript, n8n, LangGraph or the vendor tools you already pay for, connected to your APIs and internal systems.
Architecture and scope agreed in writing before the build
Built on your infrastructure or the platform you choose
Integrations with your APIs, database and internal tools
Fixed-price quote
Scoped after a call. A written quote within 2 business days of receiving your brief.
Monitoring, fixes and updates for agents already in production.
After launch, models get new versions and the APIs an agent calls change. We monitor, fix and update the agents and automations you run, including ones someone else built, for a flat monthly fee. Response times are agreed in the contract.
Monitoring, alerts and incident response
Model, runtime and dependency updates
A monthly review of cost, quality and what to change next
Flat monthly retainer
Quoted to your setup. Model and hosting usage billed at cost, never marked up.
A ranked list of what to automate, and what it would cost.
We interview the people who do the work, map how it actually flows and write up a roadmap: which jobs are worth automating, ranked by return and effort, what has to change in the process first, and a fixed quote to build them.
Interviews with the people who do the work
Process maps showing where the time goes
Opportunities ranked by return and effort
€1,200 to €1,800
Scoped after a free call. The report is yours to use with any builder.
You have a fixed price in writing before any work starts. The full process
01First
A 30 minute call
We talk through the job, the tools involved and how you will know it is working. The call is free, and you speak with the engineer who would build it.
02Within 2 business days
A written quote
Scope, timeline, a fixed price and a list of what you will own at the end. We do not bill by the hour.
03Build
Weekly demos
Each week you see the agent working on your own data. It only acts on its own once you have approved each step.
04Launch
Handover
You get a runbook, monitoring, alerts and the credentials in your name. After that, keep us on a maintenance retainer or run it yourselves.
04/About
Who builds it
Magentic Studio is led by David Silva. We also build and operate Qoren, a platform that hosts AI agents for its customers. The standards below apply to every agent we deliver.
We build Qoren
Qoren runs each customer's agents on a dedicated machine, with encrypted secrets, audit logs and hard spend caps. Done-for-you agents run there too.
Guardrails
Agents get only the access they need. A person approves anything that cannot be undone, every action is logged, and every model key has a spending limit.
In writing
The price and the scope are agreed in writing before we start, maintenance response times are in the contract, and every handover includes a runbook.
Yours to keep
Accounts, code and data are in your name. You can move it all in house or to another provider at any time.
05/Pricing
Prices
These are the prices we can publish. Anything else gets a fixed-price quote in writing within 2 business days. How we price
Scoped after a free call. The report is yours to use with any builder.
06/Self-serve
Run agents yourself on Qoren, from $39 a month.
Qoren hosts OpenClaw, Hermes and Codex agents in managed cloud environments, with ready-made templates and spend caps. You do not need Docker or a VPS. If you get stuck, you can bring us in later.
It designs, builds and runs software agents that do tasks on their own, such as answering leads, sorting an inbox, updating a CRM or writing reports. Magentic Studio agrees the scope with you, builds the agent on Qoren or on your own stack, deploys it with monitoring and spend limits, and maintains it after launch if you want us to.
How much does it cost to build an AI agent?
A done-for-you agent on Qoren starts at $2,000 per workflow, plus Qoren hosting from $39 a month. Custom agents on your own infrastructure get a fixed-price quote after a call, because the cost depends on integrations and security needs. Model usage is billed at cost. If you are not sure what to build first, the AI audit costs €1,200 to €1,800.
How fast will I get a quote?
Within 2 business days of receiving your brief, you get a written quote with the scope, the timeline, a fixed price and what you will own at the end. If something in the brief is unclear, we ask first by email or in a short call.
What is the difference between a done-for-you Qoren deployment and a custom deployment?
A Qoren deployment runs OpenClaw, Hermes or Codex agents on Qoren, our managed agent platform, so there are no servers for you to run and it is the quickest to set up. A custom deployment is built on your own cloud or the platform you choose, for teams whose data residency, security or integration needs rule out managed hosting. On the call we recommend whichever fits the job.
Can you maintain agents or automations that someone else built?
Yes. We start by reviewing what is running and fixing anything risky. After that it goes on a flat monthly maintenance retainer, with response times agreed in writing.
Who owns the agents, the code and the data?
You do. Accounts are created in your name, code is delivered to your repository, and your data stays in your systems or your own Qoren account. You can move everything in house or to another provider at any time.
Which tools and models do you work with?
On Qoren we deploy OpenClaw, Hermes and Codex agents. Custom builds use Python or TypeScript, frameworks such as LangGraph, workflow tools such as n8n, and models from OpenAI, Anthropic, Google and open-weight providers. No vendor pays us. When Qoren is the right fit we say so, and we tell you that it is ours.
Where are you based, and who do you work with?
Magentic Studio is based in Portugal and works remotely with small and mid-sized teams in Europe and the US.
Tell us what you want automated.
Book a call to talk it through, or send a short brief for a written quote. Either way, you deal with the engineer who would build it.