Local-first agent engineering · Singapore

We set up and run AI agents on hardware you own.

Curiosity Garage is a small engineering practice for teams running AI agent harnesses on machines they control. We install the harness properly, build the applications around it, and — when one dies at an inconvenient hour — join your machine securely and bring it back.

Scope a harness setup See what we do

We work on your hardware, with your keys, and hand back a written record of what changed. Already down? Start an Agent Rescue →

Your agent stack, your boundary. Our access is scoped, external, and revocable.

Services

Three things, done properly.

Every engagement runs on your hardware, with your keys, and ends with a written record of what changed.

01

Agent harness setup

Install, wire, and harden a local agent harness on hardware you control — from a laptop to a workstation to a small rack.

Model and provider wiring, with your own API keys
Memory, tools, sandbox, and scheduling configuration
Health checks, backups, and a written runbook

Scope a setup →

02

Custom applications

Applications built on top of your harness — the parts that make an agent useful to the rest of your team.

Internal tools and dashboards driven by your agent
Integrations with the systems you already run
Deployed on your infrastructure, not ours

Describe a build →

03

Agent Rescue

When a harness fails — misconfiguration, dead process, broken upgrade — we diagnose and repair it with you on the call.

Secure peer-to-peer access, no inbound ports
You watch every command as it runs
Root-cause report handed back after the fix

How a rescue works →

Agent Rescue

Access that you can watch, scope, and revoke.

A rescue means trusting someone with a connection to your machine. We designed the process so that trust is verifiable at every step — and so that nothing is left open when we're done.

01

You describe it

Tell us what broke, what changed, and what you've already tried. Logs and screenshots help.

02

Triage call

Thirty minutes to decide whether this is a repair, a rebuild, or something you can fix yourself with guidance.

03

Secure connect

You install Tailscale and share a single node with our rescue tailnet. No credentials change hands; the key expires.

04

Fix together

A live session, screen shared, narration on. Nothing runs off-screen, and you can cut access at any moment.

05

Handback

Root cause, the actual fix, and what to watch for. Access is revoked when the session closes.

No inbound ports opened on your network, ever
WireGuard end-to-end, peer-to-peer — no relay we control
Ephemeral keys that expire in 24 hours, or the moment you revoke
Credentials never leave your machine; we ask you to type them
Full command log and session transcript, handed to you
Scope agreement signed before any access is granted

Key issued → scoped session → revoked. Nothing persists after the work is done.

Approach

How we work.

A

Your hardware, your keys

We don't resell hosting or hold your API keys. Everything we build runs on machines you own and can inspect, and every credential stays in your possession.

B

Read before touching

Logs, config, and git history first. We don't reinstall, upgrade, or "clean up" a working system to make a problem look solved.

C

You keep the receipt

Every engagement ends with written findings: what was wrong, what we changed, and how to recognise it next time. No black boxes, no retainer handcuffs.

Contact

Tell us what broke.

Send a short description of the system and the failure. If it's urgent, say so in the subject line and we'll prioritise triage.

Email an Agent Rescue request Ask about a setup

Prefer to write directly? contact_us@curiositygarage.com — the address is live and monitored.

What to include: OS and hardware · harness and version · what changed last · the exact error · what you've tried.
Response: Same business day for triage. Emergencies attended to in the order received.
Terms: Scope and confidentiality agreed in writing before any access is granted.