Tools I Use To Build Every Claude Code & Codex Project

Product / Tools Reviewed (Video Overview)

This video isn’t a single product review. It’s a curated list of 29 “S-tier” AI-native tools the creator uses to build real projects with Cloud Code and Codex (AI coding agents).

The “product” being evaluated is essentially the tooling stack for shipping AI-assisted coding projects.

Overall Selection Criteria (How Tools Are Judged)

All tools on the list meet the creator’s 3 criteria:

1. Free / low cost (generous free tiers or very cheap to start)
2. Scalable (works for small demos and grows to real usage / paying customers)
3. AI-native (integrates well with coding agents like Cloud Code / Codex via things like MCPs, CLIs, and plugins)

Main Recommendations by Category (Unique Tool Points)

Coding Agents

• Cloud Code and Codex (top recommendation)
  • Described as “best value” and “most powerful” by the creator
  • Uses the desktop app and CLI
  • Preferred over other agents mentioned: Coda, Open Code, Amp
  • Reasons cited: model performance / intelligence (examples like Opus 5.5 / GPT-6 Astra), plus ease of use

Hosting & Infrastructure

• Vercel and Cloudflare (both recommended; used in production by the creator)
  • If forced to pick one: the creator leans Cloudflare
  • Shared strengths:
    • Powerful yet easy to use
    • AI-native integration (MCPs / CLIs / plugins), especially preview environments
  • Positioning: Cloudflare is framed as “all-in-one,” with additional infrastructure services
• Resource access mentioned: ainativebuilder.com/stack
  • Hosted on Cloudflare
  • Built with Astro

Websites / Web Apps

• Websites: Cloudflare + Astro
  • Goal: “slick, fast” landing/business sites with good UX
• Web apps (SaaS-style): Vercel + Next.js stack
  • Next.js, shadcn UI, Tailwind CSS
  • Claim: deep expertise isn’t required—high-level understanding is enough

Databases

Primary picks (AI-tool-friendly):

• MongoDB or Supabase
  • Rationale: strong experience with AI coding tools, including MCP servers and plugins

Other options mentioned:

• Neon (Databricks-owned)
• Tiger Data for analytics / time series
  • Uses Timescale DB (time-series DB on Postgres)

Disclosure:

• The creator previously worked at MongoDB and Tiger Data

Authentication

• Clerk (single pick)
  • Emphasis on developer experience and pre-built UI for social logins
  • Strong MCP/agent-native experience

Emails

Two email types:

• Marketing / capture emails: Kit
• Transactional / product emails: Resend and Cloudflare
  • Both described as “super AI native” with MCPs/CLIs
  • If forced to pick one for transactional: Resend (Cloudflare as an alternative)

Cloud / VMs

• DigitalOcean Droplets
  • Positioned as a simpler “plain VM” option versus some serverless patterns
  • Example use: cron jobs / always-on servers

Object Storage (Media / Large Files)

• Cloudflare R2
  • AI-native; integrates with Cloudflare MCPs/CLIs
  • Competitive pricing
  • Supports private buckets
  • Compatible with AWS S3 workflows
  • Suggested use case: image generators / infographic generators

Payments

• Stripe
  • Strong agent/AI experience (“first class”)
  • Historically well-liked by engineers
  • Recommended for subscriptions and core payment integrations

AI Models Inside the App

• Text/functionality models:
  • Anthropic (Claude API)
  • OpenAI API
• Image models:
  • Gemini (specific mention of “Nano…” series as cheap)
• Additional mention:
  • OpenAI GPT Image 2

Monitoring & Analytics

• Primary: PostHog
  • Creator has been using it for “past 1 or 2 months”
  • Agent-first workflow: claims they haven’t manually set up dashboards—asked Claude/Code/Codex to build them
  • Mentions compatibility with daily inspection of agent-generated dashboards
• Also mentioned:
  • Sentry (more monitoring-focused)
  • Others in the space (not favored over PostHog/Sentry): Amplitude, Mixpanel
• Creator’s stated approach:
  • Start with PostHog, use Sentry as an alternative

Project Management

• GitHub + GitHub Issues
  • Repo organization / task tracking
• Linear
  • Planning bigger projects
• Reasons: both described as AI-native with GitHub CLI / Linear CLI + MCP
• Value: coding agents can create/write issues in preferred formats

Pros (As Stated in the Video)

• Strong AI-native integration (MCPs, CLIs, plugins for Cloud Code / Codex)
• Cost-effective (free tiers / low-cost starts)
• Scales from demo to production
• Emphasis on developer experience (pre-built UI components, “easy to use” hosting stacks)
• Workflow speed: agents can set up dashboards/issues/tooling with minimal manual work

Cons / Limitations (Only What the Subtitles Imply)

• No explicit downsides are listed for individual tools.
• The implied tradeoff:
  • Choose stacks based on whether you’re building websites vs web apps
  • Don’t force one stack to cover all use cases

Comparisons Made

• Cloud Code / Codex vs other agents
  • Creator tried others (Coda, Open Code, Amp) but returned to Cloud Code/Codex due to model performance/intelligence and ease of use
• Hosting choice
  • “If I had to pick only one,” the creator would choose Cloudflare over Vercel (though both are included)
• Monitoring tools
  • PostHog vs Sentry (plus mention of Amplitude/Mixpanel as alternatives)
• Transactional email
  • Resend vs Cloudflare (Kit is for marketing/capture)

Numerical Ratings / Scores

• No numeric ratings or star scores are provided.

Concise Verdict / Recommendation

The creator strongly recommends adopting an AI-native, agent-integrated tooling stack (Cloud Code / Codex + the 29 listed tools) because it’s:

• Free / low cost
• Scalable
• Most importantly: it works directly with coding agents via MCP / CLI / plugin support

If you’re building production projects with AI coding agents, the stack is presented as a practical go-to blueprint.

Speakers / Contributors

• Single speaker viewpoint throughout (the creator).
• No separate speaker perspectives are included.