Layer·Domains·Developer tooling
The tools teams use to build with AI, from coding assistants to agent frameworks. Real value comes from workflow fit and reliable outcomes, not flashy demos.
Cursor
AI-native IDE with deep codebase context. Fastest-growing coding assistant by developer adoption. Agent mode and multi-file edit are the primary differentiators.
API-firstApplication
Cursor
Cold read
Cursor is an AI-native code editor built on VS Code, developed by Anysphere. It indexes entire codebases to provide inline suggestions and multi-file agent edits.
Position read
Grew rapidly through developer word-of-mouth after shipping deep repo indexing and multi-file agent mode ahead of incumbents. Now in a more competitive market as GitHub Copilot added agent features and Windsurf gained momentum. The differentiation that still holds is codebase context depth — it indexes the full repo, not just the open file. SpaceX holds a $60B acquisition option on Anysphere.
GitHub Copilot
Microsoft-backed coding assistant embedded in VS Code and GitHub. Largest installed base. Multi-model backend (GPT-4o, Claude, Gemini) launched in 2025.
PlatformApplication
GitHub Copilot
Cold read
GitHub Copilot is a coding assistant developed by GitHub, a Microsoft subsidiary. It is embedded in VS Code and GitHub and uses a multi-model backend including GPT-4o, Claude, and Gemini.
Position read
Has the largest installed base but shipped agent features after Cursor and Windsurf. The distribution advantage is durable: it ships inside GitHub and VS Code, so enterprise teams often adopt it without a separate procurement decision. Multi-model backend support gives admins flexibility, which matters in security-conscious environments.
Windsurf
Codeium's AI IDE. Cascade agentic model and Flow state are the core differentiators. Competing directly with Cursor on agent-native coding experience.
API-firstApplication
Windsurf
Cold read
Windsurf is an AI-native IDE developed by Codeium. Its Cascade agentic model and Flow state are the primary differentiators over standard code completion tools.
Position read
Codeium's AI IDE, competing directly with Cursor on agent-native coding. The Cascade agentic model and Flow state are its core differentiators. Smaller installed base than Cursor or Copilot but strong developer sentiment and significant recent investor interest.
LangChain
Builds LangChain (orchestration framework) and LangSmith (observability). Dominant framework for agent and RAG pipeline development.
API-firstOrchestrationObservability
LangChain
Cold read
LangChain is an open-source framework for building LLM applications and agent pipelines, available in Python and JavaScript. LangSmith is its paired observability and evaluation product.
Position read
The dominant framework for building agent and RAG pipelines in Python and JavaScript. LangSmith (its observability product) is used alongside the framework for debugging and evaluation. Has a large open-source community but faces competition from lighter-weight alternatives.
Cognition (Devin)
Builds Devin, positioned as an autonomous software engineering agent. Targets full-task execution rather than inline assistance.
API-firstApplicationOrchestration
Cognition (Devin)
Cold read
Cognition builds Devin, a software engineering agent that can execute multi-step coding tasks autonomously. The company is venture-backed and US-based.
Position read
Builds Devin, a software engineering agent designed for full-task execution rather than inline assistance. Devin can open PRs, run tests, and debug across a codebase autonomously. Enterprise adoption is early — the main open question is how much human oversight is still required in real production workflows.