Playbook: написание гигов

Цель: готовые тексты 7 гигов для копипаста в Fiverr публикатор. Каждый гиг содержит title, description, tiers, FAQ, tags.

Важно: этот плейбук — выводы из 7 гигов стратегии в формате готовом к publish. Если нужно менять — изменения сначала в strategy, потом здесь.

Принципы gig writing

Title = max 80 chars — Fiverr limit. Каждое слово должно продавать.
Description = 500-1200 words — Google’s ranking factor плюс buyer confidence.
Никаких emoji в title — выглядит cheap, нет в premium гигах.
Тiers — 3 штуки всегда: Basic / Standard / Premium. Не 2, не 4.
FAQ = 5-10 вопросов — каждый отвечает на real objection или вопрос.
Tags = 5 максимум (Fiverr limit). Каждый — конкретный tech term.
Project-specific language — не “я могу сделать”, а “we ship production …”.
No superlatives (“best”, “fastest”, “cheapest”) — Fiverr banned или выглядит spam.

Общее для всех 7 гигов

Brand voice

We не I — “Upready.dev team”
Production не simple / quick / basic
Modern stack-specific — Next.js 15, Supabase, Claude Sonnet/Opus, MCP, shadcn/ui
Concrete numbers где можно (N tables, N tools, N days)
Honest limitations — что НЕ входит, чтобы избежать scope creep

Универсальный FAQ (базовые 3 вопроса для каждого гига)

Q: Can you handle production code, not just demos?
A: Yes. Upready.dev has shipped production Next.js + Supabase + Claude
   apps including our own AI-powered website builder. All deliveries
   are production-ready with TypeScript, error handling, tests where
   relevant, and deployment guidance.

Q: What if my project doesn't match one of your tiers exactly?
A: Send me details via Fiverr chat and I'll create a custom offer
   matching your exact scope. Custom offers are typical for production
   projects — they let us tailor price and timeline to your specific
   needs.

Q: Do you provide support after delivery?
A: Basic tier includes 1 revision round. Standard includes 2 revisions
   and 7-day bug fix window. Premium includes 3 revisions and 14-day
   support window. Ongoing retainers are available for long-term needs.

Гиг 1 — Claude AI integration

Tier: A (Quick-turn)

Title

I will add Claude AI integration to your Next.js app

Description

**Production-grade Claude AI integration for your Next.js app.**

We're Upready.dev — a European dev team shipping production AI-native
web apps. We've integrated Claude into multiple SaaS products including
our own AI-powered site builder. This gig is for adding Claude to an
existing Next.js application with real-world concerns handled:
streaming, error handling, tool use, context management.

## What you get

**Basic ($500, 3 days)** — Chat widget on a single page
- Claude chat widget with streaming SSE
- Single-page integration
- Error handling and retry logic
- TypeScript types throughout
- shadcn/ui chat components
- Anthropic SDK (@anthropic-ai/sdk) latest version
- Prompt engineering for your use case (1 round of iteration)

**Standard ($1 200, 7 days)** — Multi-page chat with memory
- Everything in Basic
- Multi-page chat widget (accessible from any page)
- Session memory / conversation history (stored in Supabase)
- Tool use (basic: fetch data from your backend via functions)
- Context management (truncation, prompt caching)
- User authentication-aware chat (per-user history)

**Premium ($2 500, 14 days)** — Production-grade
- Everything in Standard
- Tool calling with 5+ tools
- RAG over your documentation or knowledge base
- Prompt caching for cost optimization
- Monitoring via Langfuse or LangSmith
- Eval suite for prompt changes
- Production deployment help (Vercel / CF / custom)
- 14-day post-delivery bug fix window
- 30-minute handover Zoom call

## What we need from you

- Access to your Next.js repo (GitHub or via upload)
- Anthropic API key (we'll create one if you don't have)
- Use case description (1-2 paragraphs)
- For RAG: your documents or API access

## Tech stack

- Anthropic Claude Sonnet / Opus (latest models)
- Next.js 15 App Router (or Pages Router if needed)
- TypeScript strict mode
- Supabase for persistence (if Standard+)
- shadcn/ui for components
- Tailwind CSS

## What's NOT included (priced separately)

- Building Next.js app from scratch (see our MVP Build gig)
- UI/UX design beyond component styling
- Multi-language localization
- Mobile app version
- Voice input/output

## Why Upready.dev

We ship Claude integrations that go to production — not chatbot
wrappers. We've built our own AI-powered site builder with Claude as
the core agent. We understand tool calling, streaming, context
management, and prompt caching at production scale.

## Delivery process

1. You send repo access and use case
2. We kick off within 24 hours
3. Daily updates in Fiverr chat
4. Submit for review on delivery day
5. Up to 3 rounds of revisions
6. Final delivery with handover notes

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$500	3 days	1
Standard	$1 200	7 days	2
Premium	$2 500	14 days	3

FAQ

Q: Can you use the latest Claude model (Sonnet 4.5 / Opus 4)?
A: Yes — we use the latest Anthropic API models. We specify which
   model in the delivery based on your use case (Sonnet for most,
   Opus for complex reasoning).

Q: I don't have an Anthropic API key — can you provide one?
A: We'll help you create an Anthropic account and API key as part
   of delivery. Cost of Claude API usage is separate from this gig.

Q: Will you implement prompt caching to reduce costs?
A: Yes in Standard and Premium tiers. Prompt caching can reduce
   cost by 80-90% for repeated context — we configure this for you.

Q: Do you support tool use / function calling?
A: Yes in Standard (basic tools) and Premium (advanced, 5+ tools
   with async handling).

Q: What about RAG over my documents?
A: Premium tier includes RAG with vector search. We'll recommend
   a vector DB (Supabase pgvector is common) and handle ingestion,
   chunking, and retrieval.

Tags

claude, ai integration, nextjs, llm, anthropic

Гиг 2 — Supabase RLS audit

Tier: A (Quick-turn)

Title

I will audit and fix your Supabase RLS policies

Description

**Supabase Row Level Security audit and production fix.**

Upready.dev team — we specialize in production Supabase apps. RLS
mistakes are the #1 security issue in Supabase projects. This gig
is for auditing your existing RLS policies, finding gaps, and
implementing production-grade fixes with a test suite.

## What you get

**Basic ($400, 3 days)** — Audit of 3 tables
- Written audit report (PDF)
- Identified vulnerabilities with severity ranking
- Fix recommendations per policy
- 3 tables covered
- No implementation

**Standard ($1 000, 7 days)** — Audit + fix for 5-10 tables
- Everything in Basic, for up to 10 tables
- Implementation of recommended fixes
- Test harness for permission paths
- Verification with synthetic users (authenticated, anonymous,
  service role)

**Premium ($2 500, 14 days)** — Full production RLS rewrite
- Everything in Standard, unlimited tables
- Full policy rewrite from ground up
- Multi-tenant architecture handled
- Test suite with pytest or vitest
- Monitoring via Supabase logs + alerts
- 30-day post-delivery support window
- Incident response plan document

## What we need from you

- Supabase project URL (or service role key for read-only audit)
- Table schema (we can inspect via SQL)
- Description of user roles and access patterns
- Any known issues or recent incidents

## Why Upready.dev

We've worked with Supabase in production since 2023. RLS is subtle —
it's not just "add a policy", it's about **complete coverage**
across INSERT/UPDATE/SELECT/DELETE for every role combination. We
build test suites that verify this explicitly.

## Delivery process

1. You provide Supabase credentials (read-only for audit)
2. We run automated coverage check + manual review within 48 hours
3. Report sent with recommendations
4. Implementation starts after approval
5. Test suite runs before delivery
6. Handover with documentation

## What's NOT included

- Full Supabase migration from another platform (see separate gig)
- Schema redesign beyond RLS
- Frontend changes
- Data migration / import

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$400	3 days	1
Standard	$1 000	7 days	2
Premium	$2 500	14 days	3

FAQ

Q: Can you audit without production credentials?
A: For read-only audit we need at minimum service role key or SQL
   access. For Standard/Premium (implementation) we need write
   access to a staging environment first.

Q: Do you handle multi-tenant setups?
A: Yes — Premium tier is designed for multi-tenant architectures.

Q: What's the difference between your audit and a general security
   audit?
A: Ours is Supabase-specific. We know Supabase internals, common
   mistake patterns, and how to test RLS without breaking your app.
   Generic pen testers miss Supabase-specific issues.

Q: What if I find more issues after delivery?
A: Premium tier includes 30-day support for discovered issues.
   For Basic/Standard, additional fixes can be ordered.

Tags

supabase, rls, security audit, postgresql, database

Гиг 3 — Custom AI agent (flagship)

Tier: B (Core revenue)

Title

I will build a custom AI agent with tool use and RAG

Description

**Production AI agents — not chatbot wrappers.**

Upready.dev builds custom AI agents with real tool use, RAG pipelines,
session memory, and production-grade evaluation. We've built our own
AI-powered website builder with Claude as the core agent. We know
agent design beyond "wrap OpenAI in a chat UI".

## What you get

**Basic ($1 500, 7 days)** — Single-purpose agent
- Single-purpose agent (one job, done well)
- 3-5 custom tools (function calling)
- Streaming responses
- TypeScript types
- Anthropic Claude Sonnet/Opus
- Basic system prompt + 1 round of iteration

**Standard ($3 500, 14 days)** — Multi-step agent with RAG
- Everything in Basic
- Multi-step agent with planner (up to 5 steps per task)
- Tool orchestration (tools can call other tools)
- RAG over your documents / knowledge base
- Session memory (conversation history in Supabase)
- Basic eval suite (10-20 test cases)
- System prompt engineering (3 rounds of iteration)

**Premium ($7 000, 21 days)** — Production multi-agent
- Everything in Standard
- Multi-agent architecture (specialist agents coordinating)
- Full RAG pipeline (vector DB + hybrid search + re-ranking)
- Production monitoring (Langfuse / LangSmith / custom)
- Comprehensive eval suite (50+ test cases)
- Error handling and graceful degradation
- Cost optimization (prompt caching, model tiering)
- Production deployment
- 14-day post-delivery support
- 1-hour handover call with architecture walkthrough

## Tech stack

- Anthropic Claude Sonnet / Opus (latest)
- TypeScript + Next.js 15 or standalone Node.js service
- Supabase (Postgres + pgvector + storage)
- Optional: LangGraph, Mastra, or custom orchestration
- Anthropic prompt caching
- Observability via Langfuse

## What we need from you

- Use case description (what should the agent do?)
- Sample tasks and expected outputs
- Access to data/API for tools
- Documents for RAG (if applicable)

## Why Upready.dev

- Built production AI agents including our own AI site builder
- Understand the difference between chatbot wrappers and real agents
- Experience with multi-step reasoning, tool orchestration, eval
- Not just implementation — also prompt engineering and iteration

## What's NOT included

- UI design (agents are backend-first; basic chat UI included)
- Training custom models (we use Claude API, not fine-tuning)
- Multi-language support (English default)
- Mobile native app

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$1 500	7 days	1
Standard	$3 500	14 days	2
Premium	$7 000	21 days	3

FAQ

Q: How is this different from a chatbot?
A: A chatbot wraps Claude in a UI. An agent uses tool calling to
   query databases, take actions, plan multi-step workflows. Our
   agents can, e.g., "analyze your sales data, identify the top
   3 underperforming products, draft email campaigns, and schedule
   them" — all in one conversation.

Q: Do you handle multi-agent setups (specialist agents)?
A: Yes in Premium tier. We use supervisor / specialist pattern
   where a coordinator delegates to domain-specific sub-agents.

Q: What about evaluation — how do I know the agent works?
A: Every tier includes eval suite. Premium has 50+ test cases
   covering edge cases, error handling, and regression. Evals
   run on every prompt change so quality is measurable.

Q: Can the agent access my Supabase / internal API?
A: Yes — tool calling is designed for this. We implement tools
   as TypeScript functions that the agent can invoke with
   arguments, returning structured data back to Claude.

Tags

ai agent, claude, tool use, rag, llm

Гиг 4 — SaaS MVP Build

Tier: B (Core revenue)

Title

I will build your SaaS MVP in Next.js, Supabase, and Stripe

Description

**Production SaaS MVP in 2-4 weeks.**

Upready.dev ships production SaaS in Next.js + Supabase + Stripe +
Claude. We don't do "quick prototypes" that break on load. We build
what you'd demo to investors or sell to actual customers.

## What you get

**Basic ($2 500, 14 days)** — Landing + auth + one core feature
- Landing page with copy and design
- Auth (email magic link via Supabase Auth)
- One core feature implementation
- Stripe subscription ($X/month plan)
- Production deployment (Vercel or Cloudflare)

**Standard ($5 000, 21 days)** — Full MVP
- Everything in Basic
- Multi-page app (landing + app + dashboard + settings)
- User profile and account management
- 2-3 core features
- Email notifications (transactional)
- Stripe: multiple plans, trial, cancel flow
- Admin panel for viewing users
- Production deployment with custom domain

**Premium ($8 000, 28 days)** — Production-grade
- Everything in Standard
- 3-5 core features
- AI integration (Claude or OpenAI for 1 feature)
- Advanced admin (search, filter, export)
- Background jobs (via Inngest or similar)
- Analytics (PostHog or Plausible)
- Transactional + marketing email infrastructure
- Staging environment + CI/CD
- 30-day bug fix window
- 1-hour handover call

## Tech stack

- Next.js 15 App Router + TypeScript
- Supabase (auth + database + storage + realtime)
- Stripe Billing (subscriptions, invoicing)
- shadcn/ui + Tailwind CSS
- Claude / OpenAI (if AI feature)
- Vercel / Cloudflare Pages + Workers
- Optional: Inngest, Upstash Redis, Langfuse

## What we need from you

- Product brief (1-2 pages: what, who, why)
- Feature list (ranked by priority)
- Design references (Figma, inspiration sites, or we'll create)
- Brand assets (logo, colors, fonts) or we'll use defaults
- Content (copy, images) or we'll generate placeholders

## Why Upready.dev

- Built 10+ SaaS products across 2 years
- Production-grade code — not "pretty demos"
- Own AI-powered site builder as proof of stack mastery
- Fast delivery without compromise on quality

## What's NOT included

- Extensive custom design work (beyond applying brand to shadcn)
- Mobile native apps
- Multi-language localization (English only default)
- Ongoing maintenance (separate retainer available)
- Legal / compliance (Terms, Privacy, GDPR — we can template but
  not legal advice)

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$2 500	14 days	1
Standard	$5 000	21 days	2
Premium	$8 000	28 days	3

FAQ

Q: Can you use a different stack?
A: We specialize in Next.js + Supabase + Stripe because we ship
   production code fastest with this stack. If you need PHP, Rails,
   Django, Laravel — we're not the right fit.

Q: Will the code be mine?
A: Yes. All code ships to your GitHub repo, you own it 100%.
   No vendor lock-in.

Q: What happens after 14/21/28 days?
A: Delivery is submitted via Fiverr. You test, request revisions
   (included in tier). After approval, project closes. Ongoing
   support is separate retainer.

Q: Do you handle payment integration for my users?
A: Yes — Stripe Billing is included for subscriptions. Manual
   payments, one-time purchases, invoicing also supported.

Q: Can you add AI features later?
A: Basic/Standard tiers don't include AI. Premium tier has 1 AI
   feature. For more AI work, order separately from our Claude
   AI integration gig.

Tags

nextjs, saas, supabase, stripe, typescript

Гиг 5 — MCP server для бизнес-процессов

Tier: B (Core revenue)

Title

I will build a Model Context Protocol server for your business

Description

**Custom MCP server — turn your business into Claude Desktop tools.**

Upready.dev builds MCP servers that expose your APIs, databases, and
workflows as tools Claude Desktop and Cursor can use natively. This
is a small but rapidly growing niche — MCP is Anthropic's open
protocol for AI tool integration, and we've been building with it
since launch.

## What you get

**Basic ($1 000, 5 days)** — Single-purpose MCP server
- 1 MCP server exposing 3-5 tools
- TypeScript implementation
- Tools with typed inputs (Zod schemas)
- Works with Claude Desktop + Cursor
- Production deployment (local stdio or HTTP)
- Documentation how to install + use

**Standard ($2 500, 10 days)** — Multi-tool MCP server
- Everything in Basic
- 10+ tools
- Authentication / API key handling
- Error handling and rate limiting
- Claude Desktop + Cursor + Continue compatibility
- Docker container for self-hosting
- User guide with examples

**Premium ($5 000, 14 days)** — Production multi-server
- Everything in Standard
- Multi-server setup (separate servers for different domains)
- Database integration (Postgres/Supabase/Redis)
- Streaming tool responses
- Observability (tool call logs)
- Production deployment (CF Workers, Railway, Fly)
- Published to MCP registry
- 30-day support window

## Tech stack

- TypeScript SDK (@modelcontextprotocol/sdk)
- Zod for tool schemas
- Node.js 20+ runtime
- Optional: Cloudflare Workers, Docker, Railway

## What we need from you

- Tools description (what should Claude be able to do?)
- API documentation of services to wrap
- Credentials (encrypted, stored in env vars)

## Why Upready.dev

- Building MCP servers since launch (early 2024)
- Shipped MCP servers for our own internal tools
- Understand Claude Desktop integration quirks
- Also experienced with related protocols (OpenAI tools, LangGraph)

## What's NOT included

- Training custom models
- Non-MCP integrations (use other gigs for OpenAI, LangChain etc)
- UI for your business (MCP is backend only)
- Legal / compliance review of data access

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$1 000	5 days	1
Standard	$2 500	10 days	2
Premium	$5 000	14 days	3

FAQ

Q: What is MCP exactly?
A: Model Context Protocol — Anthropic's open standard for
   connecting Claude (Desktop, Cursor, other clients) to
   arbitrary tools and data sources. It's like plugins for AI
   assistants.

Q: Will it work with Cursor / Claude Desktop / Continue.dev?
A: Yes — MCP is standardized. A well-built server works in all
   compliant clients.

Q: Can Claude access my private data via the MCP server?
A: Only if you configure it to. You control which tools are
   exposed and how authentication works. Our deliveries include
   secure credential handling by default.

Q: Do you publish to the MCP registry?
A: Premium tier includes publishing to modelcontextprotocol.io
   public registry for discoverability. If you prefer private,
   we'll skip it.

Tags

mcp, claude, typescript, api, anthropic

Гиг 6 — Workflow → Product transformation

Tier: B (Core revenue)

Title

I will turn your N8N or Make workflow into a real SaaS product

Description

**From workflow to shipping product.**

Upready.dev specializes in transforming no-code workflows (N8N, Make,
Zapier) into actual web products your customers can use. If you have
a working automation but are limited by "I need to share this with
100 users", this gig is for you.

## What you get

**Basic ($1 500, 7 days)** — Simple workflow wrapper
- Wrap existing N8N/Make workflow
- Single-page web interface
- User auth (Supabase)
- Trigger workflow via API
- Display results
- Production deployment

**Standard ($3 500, 14 days)** — Multi-user product
- Everything in Basic
- Multi-user (per-user workflows)
- Stripe subscription ($X/month)
- User dashboard with history
- Admin panel
- Email notifications
- Rate limiting
- Production deployment with custom domain

**Premium ($6 500, 21 days)** — Full product
- Everything in Standard
- Full web app (multi-page)
- Custom logic beyond workflow (programmatic fallback)
- Analytics (PostHog)
- Transactional email infrastructure
- Background jobs (Inngest)
- Webhook integrations
- 30-day support window

## Stack

- Next.js 15 + Supabase + Stripe
- N8N self-hosted or cloud API
- Make.com API (if relevant)
- shadcn/ui + Tailwind

## What we need from you

- Existing workflow (N8N/Make) exported or shared
- Product description (who is this for?)
- Brand assets or permission to use defaults

## Why Upready.dev

- Production Next.js + Supabase experience
- Know the trade-offs between no-code workflows and programmatic code
- Can advise when to replace workflow with code (performance, scale)

## What's NOT included

- Rebuilding workflow in pure code (priced separately)
- Custom design beyond brand application
- Mobile native app

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$1 500	7 days	1
Standard	$3 500	14 days	2
Premium	$6 500	21 days	3

FAQ

Q: Will you replace my N8N workflow with code?
A: Only if it makes sense (scale, performance). Often the workflow
   stays as the automation engine, and we build a product layer
   around it. Both approaches work.

Q: Can you add features beyond my current workflow?
A: Yes — Premium tier allows for custom logic extending the workflow.

Q: Do I still need N8N hosting after this?
A: Basic/Standard yes. Premium we can discuss moving to pure code
   if performance requires it.

Tags

nextjs, saas, n8n, automation, workflow

Гиг 7 — Custom internal AI tooling (Tier C flagship)

Tier: C (High-value, flagship)

Title

I will build custom AI-powered internal tooling for your team

Description

**Production AI tooling for your internal operations.**

Upready.dev builds Cursor-like internal tools powered by Claude —
but specific to your business domain. Think: sales rep assistant
that drafts personalized outreach, code review bot tailored to
your codebase style, customer support agent with access to your
knowledge base.

## What you get

**Basic ($8 000, 21 days)** — Single-team tool
- Internal web app for one team
- Claude integration (Sonnet/Opus)
- 3-5 custom tools (tool calling)
- Session memory
- Team auth (up to 10 users)
- Basic analytics
- Production deployment (private)

**Standard ($12 000, 28 days)** — Multi-team tool
- Everything in Basic
- Multi-team (up to 50 users)
- RAG over your docs / knowledge base
- Role-based permissions
- Audit log
- Integration with 1-2 existing tools (Slack, GitHub, Linear)
- Advanced analytics
- Training session for end users

**Premium ($15 000, 42 days)** — Enterprise
- Everything in Standard
- Unlimited users (within enterprise SSO)
- Multi-agent architecture
- Integration with 3-5 existing tools
- Full RAG pipeline (vector DB + hybrid search)
- SSO (Okta / Google Workspace / Azure AD)
- SOC 2 aware (not certified, but compliant patterns)
- 60-day support window
- Architecture walkthrough and handover

## Tech stack

- Anthropic Claude Sonnet/Opus
- Next.js 15 + Supabase + Stripe (optional for billing)
- Vector DB (Supabase pgvector or Pinecone)
- SSO providers (Okta, Google Workspace, Azure AD)
- Monitoring (Langfuse, Sentry)
- Deployment (AWS, CF, or private cloud)

## What we need from you

- Use case brief (what should the tool do?)
- Team size and access needs
- Existing integrations (Slack, GitHub, etc)
- Data access and security requirements

## Why Upready.dev

- Production Claude apps including our own AI site builder
- Multi-agent architecture experience
- Enterprise patterns (SSO, RBAC, audit, compliance)
- Fast delivery without compromising quality

## What's NOT included

- SOC 2 certification (pattern compliance, not certification)
- HIPAA / PCI-DSS specific audits
- Legal / compliance review
- On-premise deployment (cloud only)

Tier pricing

Tier	Price	Delivery	Revisions
Basic	$8 000	21 days	2
Standard	$12 000	28 days	3
Premium	$15 000	42 days	3

FAQ

Q: Is this like building our own Cursor internally?
A: Yes, for specific use cases. Internal AI tools that have context
   about your data, tools, and workflows. Cursor-for-sales,
   Cursor-for-legal, Cursor-for-support, etc.

Q: What about data privacy with Claude?
A: We use Anthropic API which has no training on customer data by
   default (paid API only). Sensitive data stays in your systems;
   we build clear boundaries.

Q: Can this replace existing internal tools?
A: Partial replacement, usually. AI tools augment existing tools
   rather than replace entirely. We'll advise on what to replace
   vs keep based on your ROI.

Q: Do you handle ongoing maintenance?
A: Separate retainer available after delivery. Typically
   $1 500-3 000/month for ongoing tuning, new features, model
   upgrades.

Tags

claude, ai agent, internal tools, enterprise, typescript

Checklist перед publish каждого гига

Связанные страницы

7 гигов — стратегия — обоснование tier’ов и цен
Positioning — как Upready.dev выглядит в описаниях
Playbook: scope creep — что NOT included означает в delivery
Playbook: custom offers — upscale beyond advertised tier