Loadout configures the Claude Code / Claude Cowork setup you already pay for — CLAUDE.md, skills, subagents, and a real memory layer — so the subscription you're carrying actually does the work it's capable of. Your account, your session, your data. We hand you a configured environment; you run it.
(empty)
# no project brief
# no skills configured
# no memory layer
# every session starts from zeroidentity: solo operator, real client work
skills: [ curated, 3-5 core ]
subagents: [ fit to your workflow ]
memory: persistentillustrative — structure only, not a real client's file
§1 · What it is
Anthropic sells you the model and the compute. Nobody sells you the configuration — the part where Claude actually knows your workflow, remembers your projects, and has the right skills wired up instead of re-explaining your business every session.
§2 · Capabilities
Not a feature-card pitch — the actual components, what ships empty by default, and what Loadout installs.
| Component | Default | Loadout-configured |
|---|---|---|
| CLAUDE.md | blank — every session starts from zero | A real project brain.One file that tells Claude who you are, how you work, and what "done" looks like — instead of a blank default every session starts from zero on. |
| Skills & subagents | none installed | Built for your actual work.Curated from the parts of the free ecosystem that fit you, not the whole 100+ item pile — installed, tested, and pruned to what you'll actually use. |
| Memory / data layer | context evaporates each session | It remembers last week.A working memory/data scaffold so context, decisions, and past output persist between sessions instead of evaporating. |
§3 · Who it's for
§4 · Fence
Loadout sells packaged configuration and a setup/upkeep service — files, skill definitions, memory scaffolding, and setup labor delivered to you and installed into your own Claude account, running on your own subscription. Nothing is hosted on our infrastructure. Nobody but you ever logs into your account — that's the bright line in Anthropic's own Consumer Terms, and Loadout is built to sit clearly on the compliant side of it. If what you actually want is someone else's servers running your agents around the clock, that's a different, adjacent product — not this one.
§5 · Pricing
Setup fee plus a light monthly retainer that keeps the config current as your work changes. [Pro $20/mo, Max from $100/mo, Team Premium seat $100–125/mo — sourced, Anthropic pricing page]
| Tier | Setup (one-time) | Retainer | Fits | Includes |
|---|---|---|---|---|
| Starter | $250 | $25/mo | Pro subscribers [$20/mo, sourced] | CLAUDE.md + 3-5 core skills, basic memory scaffold, 1 setup call, docs |
| Growth | $600 | $60/mo | Max entry [from $100/mo, sourced] | Above + custom subagents for your real workflow, quarterly config review |
| Pro / Team | $1,500 | $150/mo | Team Premium seat [$100-125/mo/seat, sourced] | Above + multi-user config consistency, new-hire onboarding doc, monthly check-in |
§6 · Demo
A scripted side-by-side: one Claude session running on the default setup, one running on a Loadout-configured environment, doing the same real-shaped task. Both scenarios use synthetic client data.
Two side-by-side session transcripts plus a memory-layer walkthrough, stepped scenario by scenario. Every scenario is clearly labeled SIMULATED and uses illustrative synthetic data only — not a real client, not a real transcript.
Open the simulated demo →§7 · Limits & FAQ
Docs-style, not sales-style — including the one where the honest answer used to be "I just downgraded instead."
The free templates are the raw materials, not the finished job —
davila7/claude-code-templates and similar repos give you components; they
don't know YOUR workflow, YOUR data, or which of the dozens of skills in
awesome-claude-code (51,220★) actually apply to what you do. The
install/curation/maintenance labor is the product, not the config files themselves.
Fair, and that's the right move if the problem really is "I don't need this much capability." But if the honest answer is "I have capability I'm not using well," downgrading trades the symptom for a permanent ceiling — you'll hit the same wall on a cheaper plan sooner. This offer is priced to be cheaper than the gap between what you're already paying for Max ($100+/mo) and what you'd get from a cheaper tier — worth testing which one is actually true before downgrading.
[named pattern from a real public forum post, not a hypothetical]Yes with an important boundary, stated plainly: nobody but you ever logs into your account. Anthropic's Consumer Terms bar account/credential sharing and reselling access to their Services — this offer does neither; it delivers configuration artifacts and setup labor, installed by you, into your own account.
"You may not share your Account login information, Anthropic API key, or Account credentials with anyone else. You also may not make your Account available to anyone else." — Anthropic Consumer Terms §2, quoted directly[our working legal read, not a legal guarantee — see risk notes in the internal appendix]
It's priced against agent-orchestration labor, not a $100-200/hr traditional dev-shop rate — that's the actual reason the fee can sit below typical agency benchmarks and still carry margin.
[derived — not a vendor quote]The retainer includes a periodic review specifically because agent config drifts as your work changes — Growth and Pro tiers both include a recurring check-in, not a one-and-done handoff. If your needs grow into wanting managed, always-on hosted infrastructure rather than a config you run yourself, that's a different, adjacent offer (PITCH-014/C16) — this stays a config+service product by design, not a bait-and-switch into hosting.