Frequently asked questions

These are the questions people ask before and after they install the apiary, grouped by topic: the basics, privacy and data, cost and performance, and setup and switching.

#The basics

#What is the apiary in one sentence?

A suite of small local programs that give your AI coding agents one shared, lasting memory, so what one of them learns is remembered by all of them, across sessions, tools, devices, and your team.

#Do I have to install all five products?

No. One command installs honeycomb, nectar, doctor, and hive together, and they work out of the box. Queen, the cloud fleet orchestrator, is a separate layer you add only if your setup grows past a single machine.

#Do I need to know databases or SQL?

No. You install with one command, sign in with a click in the dashboard, and use plain commands like honeycomb remember and honeycomb recall. The technical machinery is hidden behind the dashboard.

#Which coding assistants work with the apiary?

Claude Code, Cursor, and Codex are fully supported today. Hermes, pi, and OpenClaw are in progress. A memory written from one is recalled by the others, so switching tools mid-project does not cost you your context.

#Privacy and data

#Where does my data live?

In your own Deeplake store, which you control and can host in your own cloud account. The local daemons are the only things that connect to it, and companies, teams, and projects are separated at the storage layer, not just inside the app.

#Can other people or teams see my memories?

No, unless you choose to share. The default in a team leans private: an assistant sees less, not more, until you widen sharing on purpose. Different companies and teams never see each other's data at all.

#Does the apiary send my code or prompts anywhere?

Almost nothing leaves your machine. The only outbound traffic is sign-in with Deeplake and, optionally, anonymous usage counts that never include your code, prompts, memories, file paths, or names. You can turn that off entirely with DO_NOT_TRACK=1.

#Are my API keys and secrets safe?

Yes. Secrets are stored separately from memory, encrypted, and never shown to an assistant. An assistant can cause a secret to be used, for example to call a service, without ever seeing its value.

#Can I stop it from recording for a sensitive session?

Yes. Put the apiary in read-only mode for that session. Recall still works, but nothing new gets written, and your assistant keeps working normally either way.

#Cost and performance

#Does it cost money to run?

Everyday memory, capture, recall, and the session briefing do not need their own AI model or API key. A couple of optional extras (turning sessions into skills, and the periodic self-tidying pass) use a model and are opt-in, so you decide when to spend anything.

#Does it slow my assistant down?

No. Recording is cheap and happens out of the way, and if anything ever hiccups, your assistant keeps working normally rather than stalling on it. The session briefing is deliberately small.

#Do I need an internet connection?

You need to be signed in to reach your store. The optional search-by-meaning feature runs on a small model downloaded once to your own machine, not a cloud service, so it works offline once warmed up.

#Is it free? What license is it under?

The apiary is released under the GNU Affero General Public License, version 3.0 or later. You can read the source, run it commercially or privately for free, and audit exactly what it does, as long as you keep the license notices and share your source if you run a modified version as a network service.

#Setup and switching

#What happens if I already use Hivemind?

Honeycomb and Hivemind share one sign-in, but running both at once is not supported. The dashboard detects an existing Hivemind install and offers to move you over cleanly, usually without needing to sign in again.

#Can I use the apiary across multiple machines?

Yes. Sign in on each machine with the same account, and a memory captured on one is available on the others. When a fleet gets large enough to need central identity and reporting, that is what queen adds.

#What keeps everything running if something crashes?

Doctor, the watchdog, is installed by default alongside the rest of the suite. Your operating system keeps doctor alive, and doctor keeps the other daemons alive. Skip it with --no-doctor if you would rather manage that yourself.

#How do I update or remove it?

Run honeycomb update to update the command, the daemons, and the assistant bundles together. Run honeycomb uninstall to reverse only the changes the apiary made to your tools, leaving everything else untouched.

#Common questions

#What is the apiary in one sentence?

A suite of small local programs that give your AI coding agents one shared, lasting memory, so what one of them learns is remembered by all of them, across sessions, tools, devices, and your team.

#Do I have to install all five products?

No. One command installs honeycomb, nectar, doctor, and hive together, and they work out of the box. Queen, the cloud fleet orchestrator, is a separate layer you add only if your setup grows past a single machine.

#Do I need to know databases or SQL?

No. You install with one command, sign in with a click in the dashboard, and use plain commands like honeycomb remember and honeycomb recall. The technical machinery is hidden behind the dashboard.

#Where does my data live?

In your own Deeplake store, which you control and can host in your own cloud account. The local daemons are the only things that connect to it, and companies, teams, and projects are separated at the storage layer.

#Does the apiary send my code or prompts anywhere?

Almost nothing leaves your machine. The only outbound traffic is sign-in with Deeplake and, optionally, anonymous usage counts that never include your code, prompts, memories, file paths, or names. You can turn that off entirely.

#Does it cost money to run?

Everyday memory, capture, recall, and the session briefing do not need their own AI model or API key. A couple of optional extras use a model and are opt-in, so you decide when to spend anything.

#Does it slow my assistant down?

No. Recording is cheap and happens out of the way, and if anything ever hiccups, your assistant keeps working normally rather than stalling.

#Which coding assistants work with the apiary?

Claude Code, Cursor, and Codex are fully supported today. Hermes, pi, and OpenClaw are in progress. A memory written from one is recalled by the others.

#What happens if I already use Hivemind?

Honeycomb and Hivemind share one sign-in, but running both at once is not supported. The dashboard detects an existing Hivemind install and offers to move you over cleanly, usually without needing to sign in again.