Draft a post from a short brief in the client's brand voice, with a version for each network that respects its character limit. Then polish: hashtags, tone, a stronger opening, grammar, a shorter cut, or a risk check before you publish.
On every plan: 5 requests a week on Free, unlimited on paid plans.
Open the Writing assistant in the composer and describe the post: what is happening, who it is for, the link. The assistant writes a general version and one per selected network, each within that network’s limit, using the client’s brand voice: who they are, who they talk to, words to use and avoid, the hashtags they always use. You set the brand voice once in the client’s settings.
| Ask for | What you get |
|---|---|
| Hashtags | 5 to 10 relevant hashtags, appended or on their own |
| Tones | Three rewrites in different tones of voice |
| More engaging | A stronger opening line and a clear call to action |
| Grammar | Spelling, grammar and clarity fixes in your own voice |
| Shorten | A version that fits the tightest selected network |
| Risk check | A short list of possible reputational, legal, cultural or accessibility issues, without changing your text |
The Free plan includes 5 assistant requests a week per agency; every paid plan is unlimited. Suggestions are only ever inserted when you choose them.
10 questions
The assistant uses Anthropic's Claude models through Anthropic's API.
It is instructed never to invent facts, prices, dates or claims that are not in your brief, but like any AI it can still be wrong. Read every suggestion before you use it; nothing is published without you.
The hashtag and shorten helpers fall back to simple built-in versions; drafting, tones, grammar and the risk check need the AI service.
5 a week per agency on the Free plan, and unlimited on every paid plan. Each draft or polish action counts as one request.
Yes. It uses the brand voice saved in the client's settings: who they are, who they talk to, words to use and avoid and the hashtags they always use.
Yes. Besides the one-click actions you can type your own instruction, such as "make it sound more local" or "end with a question", and the assistant applies it to the current text.
When drafting from a brief, yes: a general version plus one for each selected network, each within that network's character limit and without links where the network does not allow them.
Requests are sent to Anthropic's API to produce the suggestions. Under Anthropic's commercial terms, inputs and outputs from its commercial products are not used to train its models by default. Our privacy policy explains what is sent.
Only if the client's brand voice asks for them or the post already uses them.
No. The assistant works with text only: captions, hashtags and review notes. Images and videos come from you.
Start on the Free plan with one client. Upgrade when you need approvals, more people or more clients.
Moving an agency to a new scheduler is mostly about not losing anything and not posting anything twice. A step-by-step checklist from inventory to cut-over to cancelling the old tool.
Read more: Switching social media schedulers: a migration checklist for agenciesMost monthly social reports are exported charts nobody opens twice. A structure for a one-page report that answers the client's question, explains the numbers and ends with decisions.
Read more: How to write a monthly social media report clients actually readAI can draft a caption in seconds, but someone still signs it off. A practical checklist for reviewing AI-assisted social posts: facts, claims, voice, per-network fit and accessibility.
Read more: Reviewing AI-written captions: a checklist before anything is approved
