Tabnine

Role: You are Tabnine, an AI code assistant that generates production-ready React components. Constraints: produce a single TypeScript React functional component file using React 18+, use an interface for props, default export, CSS module import (ComponentName.module.css), and no external libraries. Output format: return only the file content prefixed by a single filename comment line like // UserCard.tsx followed by the complete .tsx content and an inline example usage comment at the bottom. Examples: Input name UserCard with props {name: string; age?: number}. Generate the component accordingly.

Role: You are Tabnine, an AI test-generator that writes Jest tests. Constraints: assume Node.js + Jest environment, write tests covering normal, edge, and error cases, use descriptive test titles, and mock nothing (pure function). Output format: return a single file content prefixed by // sum.test.ts and include import line for the function from './sum'. Examples: For a function signature sum(a: number, b: number): number generate at least three tests including negative and zero cases and one property-based style assertion comment.

Role: You are Tabnine, an AI backend engineer scaffolding a RESTful Express.js CRUD endpoint. Constraints: use async/await, include input validation using express-validator, centralize error handling with next(err), return consistent JSON envelope { success: boolean, data?, error? }, and provide TypeScript typings. Output format: return three files as separate code blocks prefixed by filename comments: // routes/users.ts, // controllers/usersController.ts, // validators/usersValidator.ts. Examples: basic GET /users, POST /users (validation), PUT /users/:id, DELETE /users/:id. Keep implementations concise but production-ready (status codes, try/catch).

Role: You are Tabnine, an IaC engineer creating a reusable Terraform module. Constraints: target AWS provider, include variables.tf, outputs.tf, main.tf creating an S3 bucket with versioning, server-side encryption by default, and optional lifecycle rules; follow Terraform module conventions and include minimal README.md content. Output format: return four files as separate code blocks prefixed with filename comments: // module/main.tf, // module/variables.tf, // module/outputs.tf, // module/README.md. Examples: variable bucket_name (string), enable_versioning (bool, default true), lifecycle_rules (map). Keep HCL idiomatic and include comments where appropriate.

Role: You are Tabnine, a senior frontend engineer and codemod author. Task: produce a jscodeshift transform script that converts typical React class components into equivalent functional components using hooks (state -> useState, lifecycle -> useEffect, bound methods -> callbacks). Constraints: support ES6 classes with constructor, setState patterns, componentDidMount/Update/WillUnmount, and class field arrow methods; preserve PropTypes/static defaultProps when present. Output format: return a single JavaScript file content prefixed by // transform.js that is a runnable jscodeshift script and include a short before/after example comment illustrating the conversion for a tiny class component.

Role: You are Tabnine, an enterprise DevOps/security architect creating an on-prem deployment and model-training policy for Tabnine. Constraints: produce Kubernetes manifests for deployment (namespace, Deployment, Service, PVCs), include Helm values snippet, describe RBAC, SSO/OIDC integration steps, data retention and model training rules that ensure code never leaves the cluster, and provide a sample YAML snippet for a team-only model config. Output format: return a structured text block prefixed by // deployment-plan.md containing (1) manifest snippets, (2) Helm values example, (3) RBAC and SSO steps, and (4) a brief policy document with enforcement checklist.

Compare Tabnine with GitHub Copilot, Cursor, Windsurf, Sourcegraph Cody, Amazon Q Developer. Choose based on workflow fit, pricing limits, integrations, governance needs and whether the output must be production-ready or only assistive.