Marketers use Claude constantly — for copy, briefs, positioning, emails, social. But every session is a blank slate. You open Claude, and it has no idea who your client is, what their brand voice sounds like, who their audience is, or what you already tried last quarter.
So you explain it all again. Every time.
If you're in-house, that's one brand. If you're an agency or freelance marketer, it's three to ten clients — each with their own voice, positioning, audience, and history. The re-briefing adds up fast.
The context Claude needs to produce good marketing copy isn't generic — it's specific. Without it, you get generic output: technically correct, structurally sound, and completely missing the brand.
With it, Claude writes like it knows the brand. The difference is entirely in what context you give it, and how fast you can give it.
Currently, most marketers paste the same boilerplate at the start of every session:
Stash makes that one call: context(). Claude gets everything it needs, loaded from your record store, without you typing it out again.
| Category | What you store | Why it helps |
|---|---|---|
| Brand voice | Tone, vocabulary, what to avoid, exemplary sentences | Claude writes in the right register from the first message |
| Audience | Primary segment, motivations, pain points, what they don't want to hear | Copy lands on the right people, not a demographic abstraction |
| Positioning | How you're positioned vs. competitors, what makes you different | Messaging stays on-brief, not generic category copy |
| Campaign history | What ran, what worked, what flopped, key learnings | Claude builds on what you know, not what sounds good in theory |
| Active briefs | Current campaign, deadline, deliverables, constraints | Claude knows the context without a preamble every time |
If you manage multiple clients, Stash's collection structure maps directly to your book of business. Each client gets their own collection: brand context, audience, positioning, campaign notes.
You call context(expand=["[ClientName]"]) and Claude gets everything relevant to that client — not the whole database, just their records. Token-light, fast, clean.
Now you open Claude, call context(expand=["[RetailClient]"]), and ask:
Claude writes three options calibrated to the brand voice, the audience, and the specific learning from your spring campaign. Not because it's smarter — because you gave it the right context.
If you're managing 4–6 clients and using Claude to draft, review, and iterate on copy — you'll hit 50 free queries in your first two weeks. The math is obvious: that's 8–12 queries per client, which covers one campaign cycle.
The upgrade question is: is Stash saving more than £8 worth of time per month? For anyone using Claude seriously for marketing work, the answer is yes before the end of the first month.
Stash isn't:
It's the working memory layer: the contextual knowledge that makes Claude useful for your work rather than for marketing work in general.
Two steps:
https://app.stashlite.com/mcpFree: 2,500 records / 50 queries/month. Pro (£8/mo): 25,000 records / 500 queries. No contract. Cancel anytime.
Start with your biggest client. Add a brand voice record, an audience record, and your current campaign brief. Try it for a week. One client's worth of context is enough to know whether it changes how useful Claude is for you.