Help center FAQ
Practical questions about installing, running, and troubleshooting Honeycomb, grouped by topic. For the high-level "what is it and is my data safe" questions, see the main [Frequently asked questions](faq.md). This page focuses on the hands-on side: setup, daily operation, teams, and cost.
Help center FAQ
Practical questions about installing, running, and troubleshooting Honeycomb, grouped by topic. For the high-level "what is it and is my data safe" questions, see the main Frequently asked questions. This page focuses on the hands-on side: setup, daily operation, teams, and cost.
Related:
#Getting started
How do I install Honeycomb?
Paste one command for your system: curl -fsSL https://get.theapiary.sh | sh on macOS or Linux, or irm https://get.theapiary.sh/install.ps1 | iex in PowerShell on Windows. The installer sets up anything missing, including Node, and opens a dashboard where you finish sign-in. Full per-OS steps are in Installation.
Do I need to install Node or set up a database first? No. The installer detects whether Node and npm are present and installs a pinned version if not. There is no database to configure; the helper manages all storage for you.
Which operating systems are supported? macOS, Linux, and Windows (via PowerShell). The install command is the same on macOS and Linux, with a PowerShell variant for Windows.
Which coding assistants can I connect?
Six, at once: Claude Code, Cursor, Codex, Hermes, pi, and OpenClaw. Run honeycomb setup and it wires every one it detects on your machine. A memory written from one is recalled by the others.
How do I verify the install worked?
Run honeycomb status. It reports whether the helper is running, whether you are signed in, and which assistants are wired. Then prove the round trip with honeycomb remember "..." followed by honeycomb recall "...".
How do I finish signing in?
Open the dashboard with honeycomb dashboard and click First time setup. The page shows a short code and opens a tab to approve it, creating a free Deep Lake account if you do not have one. You never copy a code out of the terminal.
Can I read the install script before running it? Yes. Visit get.theapiary.sh in a browser. It serves an inspect-before-piping page and publishes a checksum file so you can verify the script first.
The honeycomb command says "command not found" right after install. What now? Open a new terminal window and try again. Installing a global command does not refresh the current shell's path; a fresh shell picks it up. More detail in Troubleshooting.
#Everyday operation
Do I have to remember to save things?
No. Once your assistants are wired, Honeycomb captures the useful moments as you work and hands the right notes back at the start of your next session. Saving by hand with remember is for when you want to pin something down on purpose.
How do I save and find a note by hand?
honeycomb remember "the staging database resets every night at 2am UTC" saves a note, and honeycomb recall "staging database schedule" pulls back what is relevant. You do not need to match the exact wording.
What is the dashboard for?
It is a local web page (honeycomb dashboard) that shows your memories, the state of each connected assistant, your shared skills, a map of your codebase, and overall health. It is also where first-time setup happens and where degraded subsystems are named in plain language.
Why did recall come back empty? Usually one of three reasons: you are in a different project folder than where you saved the note, meaning-search has not warmed up yet so recall is matching words, or the note was never written because you were in read-only mode. The step-by-step check is in Troubleshooting: recall returns nothing.
How do I stop it recording for a sensitive session? Put Honeycomb in read-only mode for that session. Recall still works, but nothing new is written. Your assistant keeps working normally.
How do I update Honeycomb?
Run honeycomb update. It updates the command, the helper, and the assistant bundles together.
How do I remove it?
honeycomb uninstall reverses only the changes Honeycomb made to your tools, leaving everything else untouched. Add an assistant name to remove just that one.
#Data and privacy
Where is my memory stored? In your own Deep Lake store, which you control and can host in your own cloud account. The small helper on your machine is the only thing that connects to it. More in the main FAQ under Privacy and data.
Does installing send my code anywhere? No. The only outbound traffic is the sign-in with Deep Lake and, optionally, anonymous product-usage counts that never include your code, prompts, memories, file paths, or names. You can turn the usage signal off entirely.
Are my API keys safe during all of this? Yes. Secrets are stored separately from memory, encrypted, tied to your machine, and never shown to an assistant. An assistant can use a secret without ever seeing its value.
#Teams
How does a teammate's discovery reach me? When something reusable is published to the team as a skill, every teammate's assistants pick it up at the start of their next session. You do not copy files around; Honeycomb places them where each tool looks. See Honeycomb for teams.
Will my private notes get shared by accident? No. The default leans private: an assistant sees less, not more, unless you widen sharing on purpose. Promoting something from "just mine" to "the whole team's" is always a deliberate, recorded action.
How does memory stay separated across projects and teams?
Honeycomb organizes memory by company (org), team (workspace), and the specific repository or folder you are in (project). It scopes to the folder you are working in automatically, so a note from one repo does not surface while you work in another. Switching company or team is a deliberate choice with honeycomb org switch and honeycomb workspace use.
I switched team and my queries look wrong. Why?
Your active team changed but your sign-in was still bound to the previous one. Honeycomb detects this and re-mints a corrected sign-in at the start of a session. Start a new session, and confirm with honeycomb org list and honeycomb workspace list. Detail in Troubleshooting: account and switching.
Can I use Honeycomb across more than one machine? Yes. Sign in on each machine with the same account, and a memory captured on one is available on the others.
#Cost
Does the everyday memory cost money in AI usage? No. Capturing, recalling, and the start-of-session briefing do not require their own AI model or API key. Two optional extras can use a model: turning sessions into summaries and skills, and the periodic self-tidying loop. Both are opt-in, so you decide when to spend. See the main FAQ under Cost and performance.
Does meaning-search cost money to run? No. Meaning-search uses a small language model that runs locally on your own machine, downloaded once, not a cloud service. There is no per-query charge.
Does it slow my assistant down? No. Recording is cheap and happens out of the way, and if anything hiccups, your assistant keeps working normally. The start-of-session briefing is deliberately small.
Is Honeycomb free to use? Honeycomb is licensed under the GNU Affero General Public License v3.0 or later. You can use it commercially or privately, free of charge, as long as you keep the license notices and share your source if you run a modified version as a network service.
Still stuck? Work through Troubleshooting, or start fresh with Getting started.