Everyday tools for Prompt Engineers (that you’ll actually use)

Being a prompt engineer doesn’t mean you spend your days in a lab wearing a white coat and whispering sweet nothings to a chatbot (although, let’s be honest, sometimes it does). It means you use smart tools that make you faster, sharper, and better at coaxing useful, clever, and downright magical outputs from AI models. Whether you’re just getting started or knee-deep in LLM land, here are the everyday tools you actually need.

1. ChatGPT (or Claude, or Gemini)

What it is: These are foundational LLMs where prompt engineers do the bulk of their experimentation, prompt testing, and iteration.

Why you’d need it: You’re literally engineering prompts. This is your lab, your playground, your coworker who never sleeps.

How you use it (real world example): Writing a prompt that gets ChatGPT to generate product descriptions in five tones of voice. Testing how Claude handles long-form reasoning. Comparing Gemini’s hallucination tolerance. It’s your job to figure out how each model behaves.

2. OpenAI Playground

What it is: A browser-based interface for testing OpenAI’s models (like GPT-4) with detailed settings.

Why you’d need it: Because sometimes you want to fine-tune temperature, max tokens, or system prompts without writing code.

How you use it: You’re testing how GPT-4 responds to a user input with different temperatures to see if it becomes more creative or stays conservative. It’s where you tweak settings like a mad scientist.

3. Notion (or Obsidian)

What it is: A digital note-taking and organization tool.

Why you’d need it: To keep track of prompts, failures, prompt libraries, best practices, outputs, and your eventual AI overlord gratitude journal.

How you use it: Organizing prompts by use case, tagging prompts for tone, style, or model, and logging results from prompt testing. It’s your second brain.

4. VS Code or Jupyter Notebook

What it is: Code editors and environments where you can run scripts, APIs, and test chains.

Why you’d need it: For engineers who move beyond the GUI, these are your go-to platforms for building pipelines, chaining prompts, or working with LangChain, FastAPI, etc.

How you use it: Writing a Python script that queries GPT-4 using an API, stores the response, evaluates the output, and logs it.

5. LangChain or LlamaIndex

What it is: Frameworks that help you build context-aware, multi-step, agentic AI applications using LLMs.

Why you’d need it: If you want to connect prompts to tools, search, documents, or create agents that reason and act, this is how you do it.

How you use it: You create a chatbot that can search a product catalog, pull the most relevant data, and generate an answer that cites its sources. Boom. Enterprise.

6. Midjourney / DALL·E / Runway

What it is: Generative AI art platforms.

Why you’d need it: If you’re doing multimodal prompt work (text-to-image or video), these are essential.

How you use it: You create image prompts like “cheeseburger, french fries, and soda on a red tray” and test how different camera styles or aesthetic terms change the results. Or you turn storyboards into motion sequences in Runway.

7. Google Sheets or Excel

What it is: Spreadsheet tools. Yes, they still matter.

Why you’d need it: For organizing prompt variables, A/B testing outputs, comparing results across models, or tracking content QA.

How you use it: You test 50 variations of a product description prompt and log tone, clarity, and hallucination rate. If it sounds boring, congrats—you’re doing actual work.

8. Postman or Insomnia

What it is: API testing tools.

Why you’d need it: For engineers using model APIs, these let you test endpoints, payloads, headers, and more without building a full UI.

How you use it: You hit the OpenAI API with structured JSON requests, tweaking the prompt or model ID, and seeing the live output response.

9. Figma / Canva

What it is: Design tools for wireframes, mockups, and content layout.

Why you’d need it: If your prompts generate images, UI content, or anything visual, you need a way to test how it lands in the actual product design.

How you use it: You generate onboarding copy in ChatGPT, drop it into a Figma flow, and see if it feels too wordy in the mobile layout.

10. Your brain (on caffeine)

What it is: The original, irreplaceable tool.

Why you’d need it: Because prompt engineering is more art than science. You’ll need intuition, curiosity, and a weird sense of humor to find what works.

How you use it: You experiment. You iterate. You fail, you tweak, you retry. And then you go write a blog about it.

Want to learn more about the skills that pair with these tools? Read So You Wanna Be a Prompt Engineer?