12 Client Codebases. 12 Different Stacks. One Claude Code Setup.
Stop running /init on every new project. 463+ developers run one command per client and start shipping in their stack from prompt one.
Built specifically for Claude Code · Skills, subagents, CLAUDE.md, hooks
30-day money-back guarantee · One-time payment · Lifetime updates
The handoff tax you pay every Monday
You opened a new client repo this morning. You ran /init and got a 200-line generic CLAUDE.md. Forty-five minutes later, you're still tuning it. Adding the Drizzle convention. Documenting the auth flow. Explaining that yes, it's Next.js 15 with App Router, not Pages Router from Claude's training data.
That's the handoff tax. Multiply it by your active client count: 8 clients × 1 hour setup = a billable day per cycle you don't bill for.
It compounds. Every time you switch back to a project after a week away, Claude starts cold again. The patterns you taught it last sprint? Half of them got de-weighted as your CLAUDE.md grew past 600 lines and Anthropic's context system started ignoring sections it judged irrelevant.
Your hourly rate exists because context-switching is expensive. The math should also apply to the Claude Code setup you do for every client.
30-day money-back guarantee · One-time payment · Instant access
Every client has a different “almost right”
Client A uses Drizzle. Claude writes Prisma code.
Client B uses CSS Modules. Claude reaches for Tailwind.
Client C is on Rails 7 with Solid Queue. Claude suggests Sidekiq.
Client D builds with React Server Components. Claude hallucinates client components with 'use client' everywhere.
Each project's “almost right” is a different almost-right. A single CLAUDE.md tuned for Client A actively makes things worse for Client B. A skill from awesome-skills-marketplace written for Pages Router is harmful in Client A's App Router project. There's no universal config — the whole point is that the config matches the codebase.
The fix isn't one mega-prompt. It's per-project context that travels with the project, generated from that project's actual code.
See the difference
Here's what changes when your AI actually knows your project — generic output vs code tailored to YOUR codebase.
// WITHOUT SUMMONAI
// routes/users.ts
import { Hono } from 'hono'
import { db } from '@summonaikit/db'
import { users } from '@summonaikit/db/schema'
const app = new Hono()
// Raw Hono route - skips oRPC procedures
app.get('/api/users/:id', async (c) => {
const id = c.req.param('id')
// Uses query syntax instead of SQL syntax
const user = await db.query.users.findFirst({
where: (users, { eq }) => eq(users.id, id)
})
if (!user) {
return c.json({ error: 'Not found' }, 404)
}
// No input validation, returns raw entity
return c.json(user)
})// WITH SUMMONAI
// packages/api/src/routers/users.ts
import { eq } from 'drizzle-orm'
import { z } from 'zod'
import { protectedProcedure } from '..'
import { db } from '@summonaikit/db'
import { users } from '@summonaikit/db/schema'
export const userRouter = {
getUser: protectedProcedure
.input(z.object({ id: z.string().uuid() }))
.handler(async ({ input, context }) => {
// Uses your Drizzle SQL syntax convention
const user = await db
.select()
.from(users)
.where(eq(users.id, input.id))
.get()
if (!user) {
throw new ORPCError('NOT_FOUND', 'User not found')
}
return { user }
}),
}// routes/users.ts
import { Hono } from 'hono'
import { db } from '@summonaikit/db'
import { users } from '@summonaikit/db/schema'
const app = new Hono()
// Raw Hono route - skips oRPC procedures
app.get('/api/users/:id', async (c) => {
const id = c.req.param('id')
// Uses query syntax instead of SQL syntax
const user = await db.query.users.findFirst({
where: (users, { eq }) => eq(users.id, id)
})
if (!user) {
return c.json({ error: 'Not found' }, 404)
}
// No input validation, returns raw entity
return c.json(user)
})Click “With SummonAI” to see the transformation.
Starting at $99 · One-time payment · 30-day guarantee
30-day money-back guarantee · One-time payment · Instant access
One command. Every client repo. Forever.
Run summonaikit in Client A's repo. 3 minutes. SummonAI Kit reads the codebase, detects the stack, and generates a .claude/ directory tailored to that project — skills, subagents, CLAUDE.md, hooks settings.
Run the same command in Client B's repo. 3 minutes. Different stack, different skills, different subagents. Different .claude/ directory.
Same license. Lifetime. Used across every client repo you'll ever touch.
Client A — Next.js + Drizzle monorepo
Client B — Rails + Hotwire shop
Each client's .claude/ directory commits to their repo. When you come back in three weeks, it's still there. When you onboard a subcontractor, it's still there.
30-day money-back guarantee · One-time payment · Instant access
Subcontractor handoffs in 3 minutes
You hand a project to a subcontractor on Friday. Monday morning their Claude Code doesn't know the codebase. They ping you with questions. You spend an hour on Slack re-explaining what's already in the repo.
With SummonAI Kit committed to the project ( .claude/ lives in git), the subcontractor clones the repo and runs claude. The skills load. The subagents load. The CLAUDE.md loads. Their Claude Code writes code that fits — from prompt one.
You stop being the bottleneck for everyone you work with.
“Running multiple codebases through SummonAI Kit showed me patterns I'd been following for years without realizing it. The manual approach would have missed all of them.”
Rinaldo Festa
CTO at Cosmico
30-day money-back guarantee · One-time payment · Instant access
The billable hours math
Conservative estimate per active client per quarter:
- →1 hour initial Claude Code setup (
/init+ manual CLAUDE.md tuning) - →30 min per session × 3 sessions/week × 12 weeks = 18 hours of re-explaining
- →2 hours fixing wrong-stack output Claude generated despite CLAUDE.md
≈ 21 hours per client per quarter of unbilled Claude Code configuration and correction work.
8 active clients × 21 hours × 4 quarters = 672 hours per year.
At a blended rate of $100–$200/hr, that's $67,200–$134,400 of unpaid Claude Code setup work annually.
SummonAI Kit: $99 once. Used across every client. For as long as you do client work.
What developers say
Yes, you can build skills manually or directly within your agent, but doing it properly still takes time. The biggest effort on my side wasn't writing a skill ... it was doing a deep analysis of the codebase to generate enough well-targeted skills. SummonAI Kit is a time saver!
Does what it promises.
Dropped it into my project, immediately writes more consistent code.
The skill structure saves hours of manual setup.
Running multiple codebases at Cosmico and personal side projects, I was tired of re-teaching our patterns every session. SummonAI Kit just works. Analyzed our stack, generated the right skills, and now it writes code that actually fits our architecture from prompt one. Time saved is insane.
I used to spend the first 10 minutes of every session reminding Claude about our conventions and project structure. Now it just knows. Ran SummonAI Kit once, and every generated file already follows our patterns. It feels like pair programming with someone who actually read the docs.
Yes, you can build skills manually or directly within your agent, but doing it properly still takes time. The biggest effort on my side wasn't writing a skill ... it was doing a deep analysis of the codebase to generate enough well-targeted skills. SummonAI Kit is a time saver!
Running multiple codebases at Cosmico and personal side projects, I was tired of re-teaching our patterns every session. SummonAI Kit just works. Analyzed our stack, generated the right skills, and now it writes code that actually fits our architecture from prompt one. Time saved is insane.
See plans · Starting at $99 · 30-day guarantee
463+ developers · <1% refund rate · Built specifically for Claude Code
Common agency objections
Q: I have 8 active client repos. Do I pay 8 times?
.claude/ directory lives inside each project — they coexist, they don't conflict, and you don't pay extra. Onboard new clients without touching billing.Q: What happens to the .claude/ directory when my client wants the project?
.claude/ to git, the client's team gets the skills with the project — including all Claude Code users they have. If you don't commit it, they don't. Your call. Either way, your SummonAI Kit license stays with you for the next client.Q: My subcontractors aren't all on Claude Code Max. Does this work for them?
.claude/ directory artifacts following Anthropic's spec. Any tier of Claude Code reads them — Pro, Max, or the free tier. Your subcontractor on Pro gets the same skill activation as you on Max.Every client repo, the same setup. One license. Three minutes per project. Used across every engagement you'll take on for the rest of your career.
30-day money-back guarantee · One-time payment · Instant access
$99 once vs. $2,400+ in unbilled Claude Code setup hours per quarter.
Use it on every client repo, forever. No subscription. No per-seat.
30-day money-back guarantee · One-time payment · Instant access
Code
One license, every client repo
One-time • Lifetime access
Use on every client repo, forever
- Run on every client codebaseNo per-project fee, no per-seat fee
- Built specifically for Claude CodeSkills, subagents, CLAUDE.md, hooks — all generated from the client’s repo
- 100+ skills, only the relevant ones installed
- 20+ subagents, only those matching the client’s stack
- MCP supportNewAuto-detects servers in each client repo
- Subcontractors clone & inherit the config
- Lifetime updates includedUpdated 1 day ago
30-day money-back guarantee · One-time payment · Instant access
Full Stack
Bill more than just code
Chosen by 62% of customers
One-time • Lifetime access
Use on every client repo, forever
- Everything in Code
- Product Skills & AgentOnboarding flows, feature scoping, activation metrics, release notes for client deliverables.
- Marketing Skills & AgentLanding pages, ad copy, conversion flows, A/B tests, referral loops you can ship for clients.
- SEO Skills & AgentTechnical audits, structured data, comparison pages, template pages.
- 3 dedicated subagentsProduct strategist, marketing strategist, SEO specialist.
- Price locked forever
30-day money-back guarantee · One-time payment · Instant access
30-day money-back guarantee · One-time payment · Instant access