> Discover all available Postdom documentation from the machine-readable index: https://postdom.com/llms.txt

# Set the rules once. Keep the authority.

Choose how much each connected account can do: manual, per-post review, approved plans, or bounded autonomy. Humans keep approvals, policy changes, pauses, and exceptions.

The agent cannot approve itself, expand trust, or resume paused activity.

## Choose the review point.

Each connected account has one human-selected authority level. Move from manual control to bounded autonomy without giving the agent permission to promote itself.

- **L0 · Manual:** The agent creates a draft. A human chooses whether and when it publishes. Every publish stays with a human.
- **L1 · Post review:** Each agent post waits for exact human approval before provider handoff. Nothing publishes without a post decision.
- **L2 · Plan review:** Posts flow only inside one approved plan covering the accounts, UTC window, and post maximum. Review the boundary instead of every post.
- **L3 · Bounded autonomy:** Posts flow only inside the account's saved daily cap, visibility allowlist, and quiet hours. Routine work runs inside saved account limits.

## Routine work stays inside the lines.

L3 evaluates the saved account policy before work can flow. L2 evaluates the approved plan instead.

- **L3 account policy:** maximum posts per local day, allowed visibility values, and quiet hours in the account timezone.
- **L2 approved plan:** selected accounts, UTC start and end, objective, and maximum posts.
- **Current editable default:** 3 posts per day and 23:00–07:00 quiet hours.

## One request gets one boundary.

When selected accounts use different modes, Postdom applies the strictest level to the entire request. It does not split one post into partially authorized publishes.

Synthetic example:

- **TikTok:** L3 · Bounded autonomy
- **Instagram Reels:** L2 · Plan review
- **YouTube Shorts:** L3 · Bounded autonomy

- **Resolved authority:** L2 · approved plan required

## Authority stays human.

The agent can prepare work and start a connection handoff. A human completes provider authorization and keeps every control that changes authority.

### Agent can

- Read workspace setup, accounts, policy, and the current brief
- Start a provider-authorization handoff for a human
- Propose plans and submit supported work
- Cancel scheduled work and pause activity
- Read exact run states, outcomes, and measurements

### Human only

- Create or revoke workspace keys
- Complete provider authorization
- Approve plans and posts
- Change trust levels or account policy
- Resume paused account and workspace activity
- Resolve exceptions that require a human decision

## Stop one account or all agent work.

An account pause blocks execution for that destination. The workspace-wide pause blocks agent write activity across the workspace. Only a human can resume either one.

## Approve the boundary, not every click.

At L2, the agent proposes the accounts, UTC window, objective, and maximum number of posts. A human approves that bounded plan once; eligible posts then flow inside it.

## Return the boundary exactly.

Agents read account policy before submitting work. Approval, policy, trust, pause, and exception states return to a human with the exact resource ID.

This page has a raw Markdown twin. The exact callable contract remains in the connected tool schemas and authority documentation.

- Human authority: https://postdom.com/docs/human-authority.md
- Schedules and plans: https://postdom.com/docs/schedules-and-plans.md
- Agent contract: https://postdom.com/agent-contract.md

## Authority questions.

### Can an AI agent approve its own plan?

No. Agent credentials cannot exercise human approval routes, expand trust, change account policy, or approve their own work. The agent returns the exact plan or post state to a human.

### Can I approve a plan instead of reviewing every post?

Yes. At L2 Plan review, a human approves one bounded plan covering selected accounts, a UTC window, and a post maximum. Eligible posts can then flow only inside that approved boundary.

### What happens when connected accounts have different authority levels?

Postdom applies the strictest selected-account level to the entire request. It does not split one post into separately authorized and unauthorized publishes.

### How do quiet hours work?

At L3 Bounded autonomy, Postdom checks the account's saved quiet-hours window in that account's timezone before authorizing work. A human can edit the policy or choose a stricter authority mode.

### How do I stop agent activity?

A human can pause one connected account or activate the workspace-wide agent pause. Paused destinations block execution, the global pause blocks agent writes, and only a human can resume activity.

### Does the agent handle my social-account password?

No. The agent may create an authorization URL, but a human completes the provider's authorization flow. The agent does not receive or enter the provider password.

## Give the agent room to work.

Choose the review point, keep authority human-owned, and let permitted work move without another click.