Supervision for your coding agents: routine approvals handled, real decisions reach you. Reviewed autonomous runs with Autopilot. A shared wiki across repos.
ACTower supervises your coding agents (https://t.co/PwRjl7IMDj). Beyond watching panes:
Classification: a classifier per agent CLI approves only what it can vouch for; risky requests wait for you. Teach it your own Custom Safe Rules too.
Auditability: every request logged with its own ID, who answered it and how; export any day as JSON.
Autopilot: hand it a whole task; reviewer agents check it before you commit.
Ecosystems: a wiki your agents read and grow with every task.
Multiplexers help you see multiple agents. ACTower helps you avoid watching all of them. In our own use since July, mostly in Autopilot sessions, about 8 in 10 agent requests were answered without a person.
Here's the whole product:
How Autopilot works: https://t.co/8Pk4OpCzXb
Try it: curl -fsSL https://t.co/Th9pE7GuOj | bash
Autopilot is in Premium. New customer? 3 months free with code ACTOWER-WEB3 (card required, $25/month after, redeem by Dec 31).
Hand off a task without leaving Claude Code.
Run /autopilot-manager and say what you want. A developer agent builds it in its own copy of your repo, reviewer agents check it until every one approves, and nothing reaches your repo until you stage it.
Start with a small task.
ACTower supervises your coding agents (https://t.co/PwRjl7IMDj). Beyond watching panes:
Classification: a classifier per agent CLI approves only what it can vouch for; risky requests wait for you. Teach it your own Custom Safe Rules too.
Auditability: every request logged with its own ID, who answered it and how; export any day as JSON.
Autopilot: hand it a whole task; reviewer agents check it before you commit.
Ecosystems: a wiki your agents read and grow with every task.
Multiplexers help you see multiple agents. ACTower helps you avoid watching all of them. In our own use since July, mostly in Autopilot sessions, about 8 in 10 agent requests were answered without a person.
Here's the whole product:
Try this: Settings, turn on Save decision captures. Each entry then gets View moment, the pane as it looked when the question came in.
What Activity records: https://t.co/GtyNmkqcYO
Install: curl -fsSL https://t.co/Th9pE7GuOj | bash
Your agents worked all afternoon. What did each one ask to do, and who answered it?
ACTower's decision log keeps every request with its own ID, per pane, and who answered it: ACTower's rules, Autopilot, you in ACTower, or you in the terminal.
It stays on your machine.
Your agent says it's done. Who else read the diff?
Before you commit, run a review + revise loop in ACTower Autopilot. Your own Claude Code specialists review the staged diff, and your developer agent fixes what they flag until they approve. You relay nothing.
You still commit.
Restarting a long Claude Code session by hand: ask for a handoff, start fresh, paste it back.
ACTower does all three: past your token limit, after ten quiet minutes, it asks for the handoff, restarts Claude Code and gives it back.
Off by default. Each auto restart is logged.
Full notes: https://t.co/CtzN1A0Nwj
Upgrade the way you installed:
Homebrew: brew update && brew upgrade actower
curl: curl -fsSL https://t.co/Th9pE7GuOj | bash
Desktop app: then run actower desktop
ACTower 1.3.7 is out.
The Responder, redesigned around what needs you: your questions on top, Autopilot's folded below, busy panes one click away.
Also: a calmer Auto Restart, and Autopilot answers more on its own.
@omarmsimon Herdr, it is built for coding agents. But it shows you an agent is blocked, and you still have to go answer it. ACTower runs on top of Herdr (and tmux) and approves the safe requests itself, so only the risky ones reach you.
@hnshah Point 4 changed our days too: every agent's questions in one queue, no more hunting through terminals for the stuck one.
Point 5 is a loop for us: reviewer agents send findings back to the developer agent until they sign off. Only then does the work reach us to integrate.
Full notes: https://t.co/iu6uG7KvtL
Upgrade the way you installed:
Homebrew: brew update && brew upgrade actower
curl, or new here: curl -fsSL https://t.co/Th9pE7GuOj | bash
Desktop app: then run actower desktop
ACTower 1.3.6 is out.
Auto Restart: when an agent's conversation passes a token limit you set, ACTower restarts it fresh in the same pane. Off by default.
Also: see exactly what Codex, Cursor and Antigravity want to run or change, and Autopilot runs closer to hands-free.
Some agents in tmux, some in #herdr. Now you have two places to check who is waiting on you.
ACTower works with both: their questions land in one queue, safe ones get approved automatically, the rest wait for you.
Keep the setup you like. Where do your agents run?
Agents shouldn't start every session from scratch.
ACTower Ecosystems give agents and humans a shared organizational memory across repos and sessions.
Agents read it before each task and add to it after. Humans read the same pages in Obsidian.
https://t.co/w5GbTGz3CX