Claude by Anthropic reviews

Claude is where we think through product decisions and write and review our copy. NEXUPIRA also connects to Claude, so you can work with your NEXUPIRA documents and AI team without leaving the Claude app.

Claude reads and tags every save in magpie, so you can find anything by what's in it, and it helped me build most of the app.

Claude translates every live sentence and every chat in Sponkly. It keeps names, tone and context across 32 languages better than anything else we tested, and it streams fast enough for live calls and meetings.

honestly a big chunk of FinderGit's SwiftUI was pair-programmed with Claude. Where it earns its keep is the boring-but-risky stuff: refactors that touch a dozen files, edge cases in parsing git output I'd have missed. I still review every diff myself, it's a collaborator, not an autopilot.

Claude is where most people connect daily.dev MCP today, and it's the agent in our launch video. We also built much of this with Claude Code, and every first-party plugin in the Marketplace started as a skill we tested in Claude.

I chose Claude for its detailed feedback and ability to work through design decisions with me. While designing Vresk, it helped me refine user flows, explore different approaches, and turn rough ideas into clearer, more actionable plans.

Claude helped me turn the initial idea for HireKey into our first working MVP. It gave me a way to explore ideas, work through product decisions, and translate my recruiting experience into something people could actually use.

Claude is the best agent we've found at multi-step creative work over MCP: it reads model limits, previews credit costs and chains our tools in the right order without hand-holding. For long production workflows like a UGC ad (brief → storyboard → generation → QA) it stays on track where others drift.

We reach for Claude Code when we need deep context on a fast-changing codebase and want an engineer closely in the loop. It’s especially useful for reasoning through unfamiliar parts of the product, iterating on complex changes, and getting from an idea to a solid implementation faster.

I’m using it with remote control from my mobile, and the built-in harness and self-review checks work almost like an owner for me. It has made my life so much easier.As a founder, I can now focus more on building new product features, talking to users, and other aspects of my personal life, while it takes care of a lot of the work in the background.The one thing I’ve noticed is that Fable 5 is a very powerful orchestrator, especially when it comes to spawning and managing agents. There’s still a gap between Claude Design and Claude Code, and I think this could be an important part of the orchestration layer.Ideally, the design should be directly usable by the implementation agent, without requiring me to manually download the designs and hand them over - especially when dealing with heavy prototypes.Having a seamless flow from design -> implementation -> review would make the whole agentic development workflow much more powerful. Powerful orchestration and models, honestly - especially with Opus 4.8 and onwards. The combination makes it incredibly capable, particularly when it comes to coordinating and managing multiple agents.

Built almost entirely with Claude Code — one of us, a lot of terminal tabs, and an unreasonable number of Sundays.

My standard go to for coding, Gemini for deep research.