Use cases · Agent publishing
Your agent drafts. You approve. Nothing posts itself.
Hand your coding agent the Blowhorn CLI and it drafts, dry-runs and schedules posts for you. Every action is scripted, every write is a dry run until you confirm, and only posts you approve go out.
Sound familiar?
An agent that posts on its own
A model that decides what to click can post the wrong thing while you are looking away.
Drafts scattered across chats
Your agent drafts a post in one chat, you paste it into a composer, and the schedule lives nowhere you can pause it.
Approval by hope
If nothing in the queue stops an unapproved post, the only check is remembering to look before it goes out.
How Blowhorn runs it
Scripted actions, rehearsed first
Every action is scripted and can be dry-run first. No model decides what to click.
The same queue and limits you have
Agents get the same queue and limits you do: one queue, human-paced, with CAPTCHA and 2FA handed to you.
Only approved posts go out
An agent queues a draft and you approve it with blowhorn content approve. A caller marked as an agent is refused, so an agent can never approve its own content.
Built for agents to read
--help with examples, --json and --check-queue, so your agent can show you what it queued and what a run would do.
Or tell your agent
-
You say: "Draft the launch post from CHANGELOG.md for LinkedIn and X, dry-run it with blowhorn post --dry-run, show me the result, then wait for me to approve it."
Frequently Asked Questions
Can my agent approve its own posts?
No. Activities an agent queues wait for approval, and approving is a person's act: a caller marked as an agent is refused. You approve with blowhorn content approve, and the approval records who approved it and when.
What do I see before I approve?
--check-queue lists what a run would do, opening nothing, and --dry-run rehearses a post without sending it. With --json the output is machine-readable JSON, so your agent can show you exactly what it queued.
Which agents work with it?
Any agent that can run the Blowhorn CLI on your own computer: Claude Code, Codex, Cursor and similar. Per-agent setup lives on the agents pages. Bring-your-own-model rules per platform and profile are Coming soon.
Does it post as a bot?
No. Blowhorn types and clicks in your own browser, signed in as you, at human pace. The post goes out as you, never as a bot, and nothing is branded sent via Blowhorn.
Not a bot. You, at full volume.
Read more before you download:
Run Blowhorn from your coding agent Blowhorn for DevRel Blowhorn for founders Post from the CLI
