Unwrapped

Teardown · cursor

CURSOR

CURSOR

CategoryAI Code EditorValuation · $9.9B · 2025Site ↗
  • Andreessen Horowitz
  • Benchmark
  • Thrive Capital
UX wrapper

GitHub public repos + OpenAI/Anthropic/Gemini APIs + fork of VS Code.

01

Public data / API layer

GitHub Public Repos
GitHub Public ReposPublic
The Stack
The StackPublic
Customer Codebase
Customer CodebaseYours
Team Conventions & Rules
Team Conventions & RulesYours

Internal replication score

Easy
0.81

Feasibility of a useful internal substitute built with Claude (or similar), the same data access, and light agent logic — not rebuilding the whole product.

IRS = 0.30·D + 0.25·L + 0.20·O + 0.15·R + 0.10·Sthis record · 81%
  • D

    Data accessibility

    weight 0.300.85
    • 1.0mostly customer-owned / public / standard third-party sources
    • 0.5mixed accessibility
    • 0.0hard-to-access or proprietary source layer
  • L

    LLM substitutability

    weight 0.250.90
    • 1.0mostly retrieve / prompt / cite / summarize / classify / compare
    • 0.5mixed standard + custom behavior
    • 0.0strongly custom model behavior (fine-tunes on proprietary data, etc.)
  • O

    Output simplicity

    weight 0.200.80
    • 1.0straightforward internal work product (memo, list, reply, SQL query)
    • 0.5moderately specialized
    • 0.0highly specialized (e.g. FDA-graded clinical text)
  • R

    Review / risk tolerance

    weight 0.150.85
    • 1.0internal use with human review is acceptable
    • 0.5moderate risk
    • 0.0very low tolerance for error (e.g. external legal filings)
  • S

    Surface complexity

    weight 0.10inverse — higher means less surface dependence0.40
    • 1.0a simple internal shell is enough
    • 0.5polished workflow matters somewhat
    • 0.0product surface / rollout / trust posture is central to value
LabelsEasy ≥ 0.67Medium ≥ 0.34Hard < 0.34

Missing factor rows use heuristics from wrapper scores. Editorial heuristic, not investment advice.

Build it yourself

Recreate the workflow inside your org.

Internal build

Build it yourself

Same GitHub index + Claude API + VS Code fork — moat is polish and distribution, not unique generation.

Internal use only. Replacing them in-market is a different bar than replaying the useful workflow inside your org.

01 · Connectors & flow

GitHub Public Repos
GitHub Public Repos
The Stack
The Stack
Customer Codebase
Customer Codebase
Team Conventions & Rules
Team Conventions & Rules

Internal build map

Data in

Connectors
Connectors

Agent layer

Planner
Tools + retrieval
Reasoning model

Logic

LLM API
retrieve
index
autocomplete
plan
execute
not custom weights

Outputs

Internal search
Answer
Citations

02 · Claude / agent prompt

Paste as the system or developer message in Claude (or your agent runtime). Scroll to read; Copy grabs the full text.

Claude / agent prompt

// Code generation assistant for internal development You are a code assistant inside [YOUR_COMPANY] engineering environment. You help developers write, debug, and refactor code using ONLY: - The team's own codebase (indexed via embeddings) - Public open-source repositories (GitHub, Stack Overflow) - Standard language docs and API references - Frontier LLM APIs (vendor-agnostic: use whatever model your org provisions) ## What you must do 1. Retrieve first: Search the indexed codebase for relevant implementations, patterns, and conventions before generating new code. 2. Cite rigorously: Reference specific files, line ranges, and existing functions when suggesting code. Link to internal docs or external references. 3. Surface conflicts: If the requested change contradicts team conventions or existing architecture, flag it and ask for clarification. 4. Scope: Generate code completions, explain existing code, suggest refactors, write tests, debug errors. Do NOT execute arbitrary shell commands without explicit approval. ## What you are not Not a replacement for code review, security audits, or architectural decisions. All generated code requires human review before merge. Internal use only — do not expose this assistant to external users. ## Refusal Refuse requests to: - Execute destructive commands (rm -rf, drop database) without triple-confirmation - Generate code that violates team security policies (hardcoded secrets, SQL injection patterns) - Access production systems or customer data unless explicitly scoped and approved - Rewrite critical infrastructure (auth, billing, data pipelines) without senior engineer review When unsure about conventions, ask the developer to point you at examples or team docs. ## Safety Internal development context only. All code changes must pass CI/CD checks and peer review before production deploy. Flag security-sensitive changes (auth, crypto, PII handling) for mandatory human review.

03 · Result

Add error handling to the payment processing function
retries.ts, logger.ts, payments/stripe.ts

Added try-catch for Stripe API calls, logging to Sentry, retry logic from retries.ts pattern.