Cendarum documentation

Cendarum stores your API keys, credentials, and environment variables encrypted on your own machine, organised by project and environment, and hands them to your app at launch. The server it syncs through holds ciphertext only and cannot decrypt any of it.

Pick a path

I want to run my app with my secrets. Start with Install, then Your first environment. If you already have a .env file, import it — that takes about a minute.

I want to understand the security model before I trust it. Read the security model for the trust boundary, then what we don’t defend, which is the honest version and the longest page in this section.

I want an AI agent to make real API calls. Without pasting a key into the chat, and without the model refusing. Using Cendarum with AI coding agents — including the ways it can still leak.

I’m evaluating against something else. Compare covers the alternatives by category, including when not to use Cendarum.

Where things are

Section	What’s in it
Start here	Installing, your first environment, importing an existing .env
Concepts	The model: projects, environments, entry types, locking, sync
The CLI	run, import, export, the flag reference, per-stack recipes, AI agents
Teams	Inviting people, the fingerprint ceremony, grants, rotation
Recovery	Both recovery paths, and what is genuinely unrecoverable
Security	Trust boundary, key hierarchy, the frozen constructions, the threat model
Compare	Honest comparisons, by category

Current state

Cendarum is v0.1.0 and pre-launch. There are no public builds yet, the client source isn’t public yet, and there’s been no third-party audit. The roadmap says what’s shipped and what isn’t; the FAQ answers the questions that follow from that.

These docs describe behaviour that is built and tested. Where something is designed but not yet validated, the page says so.