Curated AI tools for builders.

Why it is useful

Cursor is worth trying when your work happens inside an existing codebase rather than a blank prompt. Its strongest use case is asking project-aware questions, making targeted edits, and iterating on implementation details while keeping the surrounding files in view. It is especially useful for developers who want AI help without leaving the editor.

Why it is useful

GitHub Copilot fits teams that already live in GitHub and mainstream IDEs. It is useful for inline completions, boilerplate, unit-test scaffolding, and quick explanations of code patterns. The value is strongest when developers treat it as a pair programmer for repetitive work rather than a replacement for review and architectural judgment.

Why it is useful

CodeRabbit is useful for teams that want another layer of review feedback on pull requests. It can highlight possible issues, summarize changes, and help reviewers focus attention. It should supplement human review rather than replace it, especially for architecture, product behavior, and security-sensitive decisions.

Why it is useful

Sourcegraph Cody is designed for codebase-aware help, particularly in larger repositories where search and context matter. It can answer questions about existing code, help navigate unfamiliar systems, and assist with changes that depend on repository structure. It is most valuable when paired with good code search habits.

Why it is useful

Claude is a strong fit for long-form reading and structured reasoning tasks. Developers can use it to summarize specifications, compare implementation options, rewrite documentation, or reason through complex product requirements. It is particularly helpful when the input is long and the desired output needs a calm, organized explanation.

Why it is useful

ChatGPT is a broad AI workspace for developers, writers, analysts, and product teams. It is useful when you need quick ideation, code explanations, prompt drafting, data interpretation, or a second pass on technical writing. Its flexibility makes it a good default assistant, though important outputs should still be checked against primary sources or runtime behavior.

Why it is useful

Perplexity is helpful when a task starts with research rather than code. It can surface sources, summarize competing explanations, and give a quick map of a topic before deeper reading. For technical work, it is best used as a discovery tool and then verified against official documentation, specifications, or source repositories.

Why it is useful

v0 is useful when the bottleneck is moving from a UI idea to a concrete React draft. It can help product engineers explore layout options, produce component starting points, and speed up early interface iteration. The generated result still needs design review and integration work, but it can shorten the blank-canvas phase.

Why it is useful

Figma AI is valuable when AI assistance needs to sit inside a collaborative design workflow. It can help designers and product teams explore interface directions, find assets, and generate early content without leaving the canvas. The output still depends on strong design judgment, but it can speed up routine exploration.

Why it is useful

Framer AI is helpful for turning a rough website idea into an interactive draft. It is especially useful for marketing pages, portfolio concepts, and product experiments where speed matters more than custom application logic. Developers may still rebuild complex pieces later, but it can accelerate visual direction and stakeholder review.

Why it is useful

Midjourney is best suited for creative exploration rather than strict production UI assets. Designers, founders, and content teams can use it to build moodboards, explore visual directions, or create concept imagery. It is strongest when the goal is inspiration and art direction, not pixel-perfect brand-compliant deliverables.

Why it is useful

Runway is useful for teams experimenting with motion, short videos, product visuals, and AI-assisted editing. It can help create early video concepts or enhance creative workflows without a full production pipeline. The best use is rapid visual exploration before committing time to polished media production.

Why it is useful

Raycast AI works well for people who prefer command-palette workflows. It keeps small AI actions close to daily operations such as rewriting text, summarizing snippets, launching automations, or transforming clipboard content. It is less about long research sessions and more about shaving seconds from repeated desktop tasks.

Why it is useful

Notion AI is useful when your team's knowledge already lives in Notion. It can help turn scattered notes into cleaner summaries, draft project documentation, and organize meeting follow-ups. Its main advantage is proximity to workspace content, which reduces the friction of moving context between a notes app and a separate assistant.

Why it is useful

OpenAI Platform is relevant when you are building AI features directly into a product. It provides model APIs and tooling for assistants, automation, analysis, and multimodal workflows. Developers should pair it with careful evaluation, safety checks, logging, and product-specific UX rather than treating model output as a complete application layer.

Why it is useful

LangChain is useful when an LLM app needs orchestration: retrieval, tools, memory-like state, chains, agents, or integrations. It gives teams building blocks for more complex AI workflows, but it also introduces abstraction. It is best adopted when the app has real orchestration needs rather than a single prompt call.

Why it is useful

LlamaIndex is a strong option for connecting documents, private knowledge, and structured data to LLM applications. It is especially relevant for retrieval workflows where data ingestion, indexing, and query behavior matter. Teams should still evaluate retrieval quality carefully with realistic user questions.

Why it is useful

Replicate is useful when you want to experiment with many models without managing infrastructure first. It helps developers test model behavior through APIs and compare options quickly. For production use, teams should consider latency, cost, reliability, and data-handling requirements before depending on a specific hosted model.