use cases, for developers
Full speed.Inside the walls.
The agent you already use, on your own machine. Edits, installs and tests go through without you; anything else waits for a yes. Free per machine, no account.
curl -fsSL https://fenzi.dev/install.sh | sh
Linux, Apple Silicon, WSL 2. Free at work too.
See how it worksMissions7
BriefingThe failures you have seenPlayer One developer
ordinary problems
Engineering answers to problems you have already had.
Each card opens the mission that answers it.
the shortcut
Allow everything.
The prompts got in the way, so the agent got full access. Nobody can say what that covers.
the queue
One agent at a time.
Nothing keeps two agents apart on one laptop, so the second task waits.
the merge
Whose change is this?
The agent edits the folder you are working in, and its changes land beside yours before you have read them.
Mission 1Room to workPlan Free
auto mode
Give it room to work. Know exactly how much.
The project copy is writable and the package indexes in the default policy are open, so edits, installs and test runs go through without you. Everything else stays shut until a line in fenzi.yaml opens it.
Claude Code starts in its manual mode and asks before each command. In auto mode it stops asking, and the policy still holds.
fenzi run claude --permission-mode autoDefault access Decision log Is dangerously skip permissions safe
time action process destination rule
10:42:05 allowed claude POST api.anthropic.com/v1/messages _provider_fenzi_claude
10:42:07 denied curl paste.example.net:443 -
10:42:31 allowed python3 GET pypi.org/simple/requests/ pypi_org
Mission 2Several projects at oncePlan Free
one sandbox per project
An agent per project, all working at once.
Start an agent in each project folder. Each one runs in its own sandbox, with that project's policy, its own log and its own staged changes, so no agent sees another project's files.
fenzi ps lists every sandbox on the machine. fenzi connect takes you back into any of them, and fenzi diff in a folder shows that project's changes.
[
{ "name": "fenzi-web-4f2a", "agent": "claude",
"folder": "/Users/dana/code/web", "running": true },
{ "name": "fenzi-api-90cc", "agent": "codex",
"folder": "/Users/dana/code/api", "running": true }
]
Mission 3Any agent, same rulesPlan Free
same walls, any agent
Pick the agent per task. The rules stay the same.
The rules bind the process, whichever model drives it, so trying another agent does not mean writing new rules. Claude Code, Codex, OpenCode and Grok today. Cursor CLI and Muse are next.
Agents run with an API key. Claude Code can also sign in with your Claude subscription. ChatGPT and Grok sign-ins are not supported yet.
Fenzi starts command-line agents. A desktop editor that opens its own terminal on the host is not inside the sandbox.
fenzi run claudeClaude Codefenzi run codexCodexfenzi run opencodeOpenCodefenzi run grokGrokfenzi run --sign-in claudeYour Claude subscription
Mission 4Review the diffPlan Free
diff and apply
Let it rewrite the module. You decide what lands.
The agent works on a copy of the project. Your folder changes only when you apply, and apply refuses a file you edited in the meantime, so your own work is never overwritten.
The sandbox stays up after the agent exits. fenzi connect picks the session up where it stopped.
fenzi difffenzi applyfenzi connect
■ 3 files changed from the agent not applied yet
edited src/api/client.ts +14 −3
new src/api/schema.json +40
deleted src/api/legacy.ts −22
? What should happen to these changes?
❯ Show the changes
Apply them to ~/code/web
Keep them for later
Mission 5New access, asked oncePlan Free
requests
New access: asked once, saved for good.
A host not in the policy is refused, and fenzi log says why. The agent files a request with its reason, and a person answers it.
Approve with --save and the host goes into fenzi.yaml after the policy check, so the next run never asks. A host you add by hand loads onto the running sandbox with fenzi policy apply.
■ 1 request from the agent fenzi-web-4f2a
7f3a91c2 crates.io:443
/usr/bin/curl “User asked for the newest serde version from the crates.io API.”
$ fenzi approve 7f3a --save
Mission 6The Mac appPlan Free
the second door
Requests, approvals and the log in one window.
Approve and add to fenzi.yaml shows the change and its policy check before it writes the host into the project's file.
On a machine signed in to a team, it forwards decisions and sends requests to the console. Every screen runs a fenzi command and shows the line it ran, so install the CLI first.
Mission 7Scripts and orchestratorsPlan Free
drive it from code
Let your scripts start the agents.
Most commands take --json and print one JSON document, with errors kept apart on standard error. Exit codes are fixed, so a script, a scheduler or an orchestrator of your own can tell a policy failure from a missing runtime without parsing text.
fenzi run and fenzi connect return the agent's own exit code.
| Code | Meaning |
|---|---|
0 | Success. |
1 | Error. The message says what failed. |
2 | The policy check failed. Nothing started or loaded. |
3 | The sandbox runtime is missing, outside the pinned range, or not connected. |
--json on doctor, ps, requests, approve, deny, log, mcp list, status, policy check and more.
Level selectThe other use cases
same walls, other seats
The same sandbox, seen from another seat.
One policy and one setup for the whole team.
All use cases and every feature with its docs page.
PlansFree and TeamSandbox hours Not billed
price
You pay for the fleet, not for sandbox hours.
$0 Per machine. No account.
Fine at work, on each developer's own laptop.
$20 Per developer a month, billed yearly.
$25 billed monthly, two developers at least. One policy, one log and one approval queue across every laptop. Machines and sandbox hours are not counted.
Price has every plan side by side.
One line to install. One command to run.
curl -fsSL https://fenzi.dev/install.sh | sh
Free per machine. No account. The model does not get a vote on its own limits.