Private beta · for developers running Claude Code & Codex agents

Run more agents,
babysit less.

Pileless is the decision inbox for AI agents. When Claude Code, Codex, or your own agent hits something it shouldn't do alone — deploy, spend, publish, reply, choose — it drops the call into one queue. You decide once — Pileless sends the answer back by webhook and the agent keeps moving.

Free · quick MCP setup · no card · runs on your key.

Free during private beta · Pro planned at $19/mo — your cloud, priority support.

your agents, blocked → 4 waiting on youtap a call to clear it ↓
DE DevOps Agent ● Blocking
Deploy v2.14.0 to production?
3 services · migration 0037_add_index.sql · CI green, rollback verified. Agent suggestion: approve — additive, no destructive ops.
✕ Reject↩ Request changes✓ Approve

The bottleneck isn't the agent. It's the handoff.

Agents can work for hours. One approval, one judgment call, one missing answer stops the whole run — and today those asks scatter across terminals, Slack threads, and tabs you already closed.

The ask arrives too late

The agent stops at 2:14am. You see it at 9. The job was "running" all night, but no work actually moved.

The ask lands anywhere

A Slack ping, a terminal you already closed, a log nobody reads. Blocking questions don't belong scattered across six tools.

You become the glue

You approve in one place, paste the decision into another by hand, then hope the agent picks back up cleanly.

More agents multiply it

One agent is manageable. Five means five invisible queues in your head, none of which tell you when they stalled.

The decision inbox for AI agents.

Every approval, review, option, and blocking question lands in one prioritized queue. You decide once; Pileless sends the answer back as structured data and the agent keeps moving.

📥

Capture every blocked decision

Agents submit approvals, options, drafts, reviews, and blocking questions into one queue — sorted by urgency, agent, and workflow.

Clear the queue fast

Approve, reject, request changes, pick one, rank options, or edit in place. Keyboard-first when you want speed, visual when context matters.

🔁

Decisions return as structured data

Every answer fires a webhook back to the agent — the decision, your notes, the chosen option, the reviewer, and an audit trail. No copy-paste.

How one decision flows

You answer once, and the workflow keeps moving.

1 · Agent hits a boundary

It reaches something it shouldn't do alone: deploy, spend, publish, delete, merge, reply, or choose.

2 · Pileless creates a decision

The agent sends the question, the context, the options, its recommended action, and a callback URL.

3 · You answer once

Approve, reject, request changes, pick one, edit in place, or add a note. One pass.

4 · The workflow resumes

Pileless returns a structured response to your callback and the agent continues right where it left off.

Open a sample pile →

Built for running more agents than you can watch.

Solo builders

Let long-running Claude Code or Codex agents pause for permission and resume the second you decide.

Agent builders

Add a human-approval layer without building queues, auth, a review UI, webhooks, and audit logs yourself.

Agencies

Review AI-generated work — copy, code, designs — before any of it reaches a client.

Teams & ops

Require human sign-off before deploys, merges, spend, account changes, or customer-facing messages.

Not another agent. Not another task board.

Pileless does exactly one job, and stays out of everything else.

It doesn't replace your agent — your agent still does the work.
It doesn't run your workflow — no models, no orchestration on your content.
It doesn't become your project tracker — the goal is an empty queue, not a board to tend.

When an agent needs a human decision, Pileless gets that decision made and sends it back. That's it.

Your agents run the work. Pileless routes the decisions.

Pileless runs no AI on your content. Your agent — Claude Code, Codex, or your own — creates the pile, waits for your decision, and resumes when the webhook returns. We never use your content for training, mining, or sale.

🧠

Plug in any agent

Wire up over MCP or the API. Your agent calls pile.create and waits for your call — any model, on your key.

🔒

Your storage, your option

On Pro and Enterprise, keep file objects in your own Cloudflare R2 bucket — they never sit on Pileless infrastructure.

🛡️

Secrets stay hidden

Detected API keys and credentials are redacted server-side before a reviewer or agent ever sees them. Least privilege by default; every access audited.

Read the full security & confidentiality brief →

Connect your agent

Private beta — MCP package access on request. Already in? Wire it up in a minute.

Claude Code

Add the Pileless MCP server to your config and your agent can pile.

Setup guide →

Codex

Same MCP server, wired into ~/.codex/config.toml.

Setup guide →

API & webhooks

Drive it directly: create piles, receive each decision on your own callback URL.

Get started →

Let your agents run — keep the calls that matter.

Keep your agents moving. Keep yourself in control. You make the call; they do the rest.

Create a test pile →