# Introducing briq

URL: https://briq.cloud/blog/introducing-briq

> Why we are building a cloud for AI agents, and what comes next.

By Sylvain Goumy, Founder of briq ([LinkedIn](https://www.linkedin.com/in/sylvain-goumy/), [X](https://x.com/goooums), [GitHub](https://github.com/goums)). 6 October 2026.

I have been a software engineer for more than 15 years. Today I build with AI agents every day.
They changed the way I work.

They also frustrate me. Every day.

## Agents work with fragile tools

Look at what an agent uses to get things done:

* **Markdown files** to remember things.
* **Skills** to know how to do things.
* **Shell scripts** written on the spot.
* **A browser**, clicking through pages made for humans.

None of this is structured. It breaks without warning. Every step costs tokens, so it is slow and
expensive. Run the same task twice, and you can get two different results.

*Figure: An agent linked to loose notes, skills, scripts and a browser, some links broken; then the same agent on a brick wall, linked to Postgres, Redis, a worker and a browser.*

## Engineering solved this long ago

Good software is built on tools that took decades to get right:

* **Databases**: structured, indexed, fast to query.
* **Clean code**: maintainable, efficient and predictable.
* **Data structures**: tailored for fast data access.
* **Libraries and frameworks**: battle-tested by millions of developers.
* **Compute**: reliable, and cheap enough to run at any scale.

A database sorts a million rows in under a second. A model cannot even fit those rows in its
context.

## Put the two together

Agents are good at understanding a goal, making a plan and writing code. Software is good at doing
the work: the same way every time, at almost no cost.

So let the agent decide, and leverage real software to do the work. The agent starts a database instead of
keeping notes in a file. It runs a script on a server instead of repeating steps in the chat. It
ships a real app instead of describing one.

You get:

* **More power.** The agent builds and runs things it could never do alone.
* **Predictable results.** The same code gives the same answer.
* **Lower cost.** The heavy work runs on CPUs, not on tokens.

*Figure: Task: find the 10 customers who spent the most in a 2-million-row export. Alone, the agent writes rows and rows of tokens and ends with an uneven ranking marked with a question mark: many tokens, slow, and the answer may vary. On briq, the agent sends one query to Postgres and gets a ranked top 10: few tokens, seconds, and an exact answer.*

## This is briq

briq gives your agent its own cloud.

With one MCP call, your agent runs a real service: Postgres, Redis, a browser, a whole app, or any
container image. Each service runs in its own isolated machine. It is billed by the second, and it
stops at the budget you set.

Your agent then works the way an engineer does. It stores data in a database. It runs code. It
tests a change on a copy before it touches the original.

## Built for agents, from day one

briq was made for agents from the first day. It was not adapted for them later.

* **The agent connects by itself.** It adds briq to its own MCP config. You sign in once and set
  its budget.
* **Every answer shows the cost.** The agent always knows what it has spent, and when the service
  will go to sleep.
* **Copies are cheap.** The agent can fork a running service, try a change on the copy, and keep
  the version that works.
* **Unused services go to sleep.** They stop on their own and wake on the next call.
* **The docs are written for agents too.** Every page is also available as Markdown.

## It's just the beginning

This first version lays the base: run a service, fork it, open it to the web, stop it, and pay
only for the time it runs.

Everything we want briq to become will build on it. Many more capabilities are coming soon. We will
write about each one here.

*Figure: A brick wall going up. The foundation is laid in orange: run, fork, expose, stop, meter. Above a dotted line, grey bricks fade as they rise, with small pictures hinting at what comes next: a chart, a rising curve, a group of agents, a brain, a clock, a plug, a globe. The top row is blank.*

## Try it

Your agent can set up briq by itself, in three steps.

1. Paste this prompt into your agent:

```text
Add the briq MCP server to your config:
- name: "briq"
- url: https://api.briq.cloud/mcp (Streamable HTTP, OAuth)
- for all my projects

Then tell me what to do next to connect and activate it. If needed, read the setup instructions for agents: https://briq.cloud/docs/setup/agents.md
```

2. Follow your agent's next steps: usually a restart, then a link to sign in.

3. Try it:

- GitHub repo: "Run hextris/hextris github repo on briq (https://briq.cloud/docs/guides/run-a-github-repo.md) and give me a link to it."
- Reporting: "Start Metabase on briq (https://briq.cloud/cookbook/metabase.md), load 'Our World in Data' energy dataset into it, generate a report on world energy generation, and share it with me." (Metabase is an open-source BI tool, like Tableau.)
- Postgres: "Start a Postgres for this project on briq (https://briq.cloud/cookbook/postgres.md) and connect the app to it."
- Web app: "Run this app on briq (https://briq.cloud/cookbook/node.md) and give me a preview link."
- Memory: "Start a SiYuan notebook on briq (https://briq.cloud/cookbook/siyuan.md), keep your notes on this project in it, and give me a link to browse them." (SiYuan is an open-source notes app, like Obsidian.)
- Planning: "Start a Plane board on briq (https://briq.cloud/cookbook/plane.md), turn our TODO list into issues, and give me a link to the board." (Plane is an open-source project board, like Linear.)

We want to know what you build with it, and what is missing. Write to us at
[hello@briq.cloud](mailto:hello@briq.cloud).

Sylvain