Great list ​@kautuk_sahni , I must try some of these tools. I have been primarily using AI for coding and debugging. You can also try Opencode and Gemini cli along with Claude code.

Have you tried to include any of the voice dictation tools such Wisprflow in your workflow.

This is a great post ​@kautuk_sahni, thank you for sharing.

Just sharing the tools I use:
           
Content Creation, Emails, Summary:

• ChatGPT: My pick for SEO strategy, finding keywords, and nailing those initial "hooks."
• Gemini, Grok: What I use for fast research.
• Claude: My choice for long-form flow and keeping the writing feeling "human."

Clean Coding:

• Claude, ChatGPT: I rely on this for the best logic, refactoring, and clean code.
• IntelliJ IDEA: My heavy-duty IDE for deep analysis, safe refactoring, and catch-all bug hunting; used for my day to day office engineering tool.
• Cursor: My favorite lightweight setup for personal projects.

Thanks for sharing this detailed breakdown ​@kautuk_sahni — I completely agree with your point that the real value comes from choosing the right AI tool for the right task rather than trying to use one tool for everything.

From my day-to-day work as an Architect working with Adobe Experience Manager, Adobe Commerce, and other Adobe ecosystem solutions, my AI stack currently looks like this:

Architecture discussions & solution design:

• ChatGPT – brainstorming architecture patterns, solution approaches, and system design discussions
• Claude – helpful for deeper reasoning around backend logic and architecture reviews

Coding, debugging & scripting:

• Claude – clean code suggestions and debugging explanations
• ChatGPT – quick scripting, API examples, and automation logic
• Cursor – increasingly useful for AI-assisted coding and code navigation

Documentation & learning:

• ChatGPT – summarizing technical documentation and exploring implementation approaches 
• Adobe Acrobat AI Assistant – reviewing long PDFs, specs, and architecture documents
• NotebookLM – useful when working with large documentation sets and knowledge bases

Research & comparison:

• Perplexity AI – quick technical research and comparing tools or vendor capabilities
• Gemini – fast exploration when looking for multiple perspectives

Like you mentioned, the key insight is that AI works best as a stack of specialized tools rather than a single universal assistant. Over time this combination has helped me significantly reduce effort in research, debugging, and documentation tasks.

Always great to learn how others are building their AI workflows — looking forward to discovering more tools from the community.

Great thread, here's my current setup — a mix of cloud tools and local stuff to keep sensitive data private.

Perplexity is my starting point for almost everything. Research, comparing approaches, understanding new concepts — it replaced traditional search for me completely.

NotebookLM is where I drop long PDFs and documents I need to actually digest. Summarization of heavy technical material is where it really shines for me.

ChatGPT & Claude are my go-to for architecture discussions and solution design. ChatGPT is great for brainstorming patterns and exploring different approaches quickly; Claude steps in when I need deeper reasoning around backend logic and architecture reviews.

Claude & Gemini also cover day-to-day reasoning and generation tasks. I tend to reach for Claude when I need structured thinking, Gemini when context length matters.

OpenCode + oh-my-opencode plugin is probably my favorite recent addition. I use it to spin up POCs fast — I let it run in the background with the plugin and come back to working code (most of the time) — and also to make sense of large, unfamiliar codebases quickly. For POCs I also switch to Antigravity depending on the task, great IDE for quick experimentation.

Buzz (Mac app) transcribes my calls. Those transcripts then go into my local Open WebUI + Ollama setup as a private RAG — nothing leaves my machine, which matters a lot for work-related conversations.

The whole stack covers research → architecture design → prototyping → code understanding → meeting memory. Each tool has its lane and they don't really overlap.

​@kautuk_sahni Thank you for sharing this. Considering the variety of AI tools and platforms available for different purposes, there isn't a one-size-fits-all solution. I use a few of them for my daily tasks, but I mainly depend on our own enterprise OpenAI's GPT-5-mini powered alternative called Pal. Additionally, Cursor is my preferred tool for AI-assisted coding.

Thanks for starting this thread ​@kautuk_sahni 🚀 ! It is really insightful to see how others are layering their AI tools.

Here's what my current stack looks like

• GitHub Copilot CLI with my custom skills for code generation, refactoring, and test creation.
• IntelliJ IDEA for full-scale development.
• Google Gemini for precise, web‑grounded answers.
• Perplexity AI for deep research.
• NotebookLM for digesting and reasoning over large document sets.
• ChatGPT for emails, summaries, and fast drafting. Also for quick scripts and configuration troubleshooting.

How I actually use AI tools in my day-to-day work (rankar.ai)