Skip to main content

AI agents

Agents are the heart of post.buzz's agent-first model. Instead of a generic "generate" button, an agent is a bounded worker assigned to one or more accounts that can plan, draft, schedule, and engage on your behalf — within limits you define, and with every decision recorded.

What an agent is​

An agent is bound to:

  • One or more channels (agent_channel) — the accounts it operates on.
  • A mode — how autonomous it is. See Modes.
  • Boundaries — how often it posts, what tone it uses, which post types it may produce, and when to pause.
  • Context — a brand profile, goals, and the account's insights.

What an agent does​

Depending on its mode and configuration, an agent can:

  • Plan a content cadence around your goals and posting windows.
  • Draft captions and select/generate media.
  • Schedule posts at recommended times, respecting posting gaps and limits.
  • Engage by replying to comments or DMs.
  • Learn from performance — what worked feeds back into future decisions.

The audit log​

Every action an agent takes is written to a decision audit log that records:

  • Action — what it did (e.g. scheduled a post, skipped a slot).
  • Decision type & reason — the category and the why behind it.
  • Insight basis — the data the decision was grounded in.

This is non-negotiable: agents are visible. You can always see what an agent decided and why, and steer it accordingly.

Boundaries that keep agents safe​

  • postsPerDay and weeklyPostLimit — hard caps on output volume.
  • minPostGapMinutes (per account) — minimum time between posts.
  • captionTone / captionLength — stylistic guardrails.
  • enabledPostTypes — which post types the agent may produce.
  • failureThreshold — if this many actions fail in a row, the agent pauses itself.
  • proposalBuffer — in approval mode, the max pending proposals queued.
  • Plan limits — monthly post caps and daily AI action ceilings still apply.

Start small​

The recommended path: begin in approval mode so the agent proposes and you review. Once you trust its judgment, raise autonomy step by step. See Modes.

In this section​