briqbriq
All postsChangelog

Connect from Claude, run long commands

Week of 12 October: briq is in Claude's directory and the MCP Registry, commands can run for hours, and databases can be reached from outside.

Sylvain Goumy

Founder, briq

17 October 2026 · 2 min read

Our first week in public. Here is what changed for you and your agents.

Connect briq from Claude

briq is now in Claude's connector directory. Open it, select Connect, then sign in to briq and choose the team and the budget Claude gets. Nothing to paste.

It works in Claude on the web, on desktop and on mobile, and in Claude Code when it is signed in with your Claude account. The connect steps on our site and in the dashboard now start with a Claude tab. Other agents keep the install prompt.

Listed in the MCP Registry

briq is in the official MCP Registry as cloud.briq/briq, and on Glama. A client that reads the registry installs briq from there. It signs in through the browser: there is no key to copy.

Run commands that take minutes or hours

briq_exec waits for a command for up to a minute. A build, a test suite or a data load takes longer. Now the agent starts it with async: true and checks on it until it ends.

briq_exec({ briq: "bq_…", cmd: ["sh", "-c", "npm ci && npm test"], async: true })
→ ex_7kq2m9x1p4ab running since 10:02 UTC

briq_exec_result({ briq: "bq_…", exec: "ex_7kq2m9x1p4ab" })
→ ex_7kq2m9x1p4ab exited 0 after 3m
  • Up to 24 hours: one hour by default, longer with timeoutMs.
  • Every line kept: each check shows the end of the output, the whole of it stays in a file.
  • Stopped on request: briq_exec_stop ends the command and every process it started.
  • No sleep in the middle: a running command keeps its briq awake.

The guide: Run a long command.

Long tool calls come back

Some MCP clients give a server 60 seconds to start answering. A briq_run waiting for a slow service could pass that limit, and the client dropped the result. Our MCP server now streams its answers, so a long call still returns.

Reach a database from outside

Ask your agent for a connection string. It gives the database a TCP endpoint, over TLS:

briq_expose({ briq: "bq_…", port: 5432, protocol: "postgres" })
→ postgres://postgres:…@db-k3m9x2p7qa.tcp.briq.run:5432/app?sslmode=require
  • Seven protocols: Postgres, MySQL, Redis, MongoDB, RabbitMQ, Neo4j and gRPC.
  • Wakes on connect: a sleeping database starts when a client connects.
  • On every plan, with no limit on the number of endpoints.
  • In the dashboard: the briq's Endpoints card, under tcp. The password stays hidden until you show it.

The guide: Connect to a database from outside.

Agents see their screenshots

briq_files_get returns a PNG, JPEG, GIF or WebP file as an image. An agent that takes a screenshot in a browser briq can now look at it. The new guide Drive a browser shows both ways to control one.

Try it

1

Add briq to Claude

In Claude: Settings → Connectors → Add custom connector.Name it briq-stg, paste this URL, then click Connect.

https://api.stg.briq.cloud/mcp

In Claude: Settings → Connectors → Add custom connector.Name it briq-local, paste this URL, then click Connect.

http://localhost:3001/mcp
2

Connect Claude to briq

Sign in, or create your account and claim your €10 of free credits.

3

Try it

Deploy briq-cloud/wall github repo on briq (https://briq.cloud/docs/guides/run-a-github-repo.md) and give me a link to it.
Deploy briq-cloud/wall github repo on briq (https://stg.briq.cloud/docs/guides/run-a-github-repo.md) and give me a link to it.
Deploy briq-cloud/wall github repo on briq (http://localhost:3000/docs/guides/run-a-github-repo.md) and give me a link to it.

Tell us what to build next at hello@briq.cloud, or ask your agent to send it with briq_feedback.

Sylvain Goumy

Founder, briq

How can we help?

Message to the briq team
hello@briq.cloud