How to Plan a Week of Social Posts with Spinach
Spinach generates social posts from the context your team already produces — meeting notes, docs, calls — and schedules them to your chosen channels. Setting up a working weekly loop takes about an hour; after that, a week of posts takes twenty minutes to review.
The value is not better copy than a human writes. It is that posting stops being a chore nobody owns on a Friday afternoon.
What does Spinach do, exactly?
It connects to sources you authorise (Google Meet transcripts, Docs, Slack, Notion, depending on plan), reads them, and proposes posts. Each proposal gets a channel, a suggested time, and usually an image.
You review in a queue, accept or rewrite, and it publishes on schedule. There is a brand voice setting that shapes tone across everything.
Pricing sits in the $20 to $50 per seat per month range depending on plan and channels; annual billing discounts it. It is aimed at teams of two to twenty, not solo creators — the setup cost only repays itself when there is context to mine.
How do you connect the right sources?
Connect fewer sources than you think you need. Four connected sources produce sharper posts than ten, because every extra source dilutes the signal about what your team actually knows.
A configuration that works for a B2B software team:
- Weekly product meeting transcript (what shipped, what broke)
- Customer support theme doc, updated weekly
- One Slack channel where the founders talk about the market
Do not connect all-hands transcripts. Ninety percent of the content is logistics and the posts will read like internal memos.
How do you set a brand voice that does not drift?
The voice field should contain two things: three phrases you would actually use, and three you never would. Negative examples work better than adjectives.
Direct, technical, occasionally dry. Say “we shipped” not “we’re thrilled to announce”. No exclamation marks. Never call a feature “game-changing”.Then, when the queue drifts, correct two posts in the editor and it recalibrates for subsequent generations. Correcting in the queue is more effective than editing the voice prompt repeatedly.## What does the weekly loop look like in practice?Monday: open the queue, which has 8 to 12 drafts for the week. Accept 4 or 5 without changes. Rewrite 2. Delete the rest.That ratio — half accepted, one fifth rewritten — took three weeks of correction to reach. In week one, expect to rewrite most of them.Check three things on every accepted post before scheduling: the claim is factual, the link works, the image is not a generic stock person pointing at a laptop. Spinach is decent on the first, weak on the second, and consistently generic on the third.## Where does it fall short?Original thinking. Posts summarise your context. If your team does not discuss the market, the posts will not contain a market opinion. The tool amplifies thinking you already did; it does not do thinking.Images. Generated visuals read as templated. Replace hero images with product screenshots or real photos — it takes one minute per post and doubles engagement in my experience.Cross-platform adaptation. A LinkedIn post and an X post from the same source often share nearly identical wording. Rewrite one of them, or the audience overlap notices.Analytics. Reporting is basic. If you need attribution, you will still stitch together platform data in a spreadsheet.## Who should use itGood fit: a two to ten person team with an active internal knowledge flow and no dedicated social hire. The setup cost is real, and the payoff is consistency.Poor fit: solo creators who do not generate much written context, and agencies managing ten client voices — brand separation across clients is more manual than the workflow can support.Also skip it if your marketing requires a strong, idiosyncratic POV. That voice comes from a person, not a queue.## Quick Summary- Spinach turns internal docs, notes and calls into scheduled social posts; setup takes about an hour.- Connect 3-4 high-signal sources, not everything — more sources dilute output.- Write the brand voice with three do phrases and three don’t phrases; correct two posts to recalibrate.- Expect a 50% accept / 20% rewrite ratio by week three; week one is mostly rewriting.- Weak on original thought and images; strongest as a consistency engine for small teams.
