AI Code Generation IDE Plugin

Part of the worlds-biggest-software-project initiative.

Context-aware code completion with project-wide understanding, delivered as an open, model-agnostic IDE plugin.

An AI coding assistant plugin for VS Code and JetBrains IDEs that combines inline completion, chat, and agentic multi-file task execution with first-class private-codebase awareness. Built for developers and engineering teams who want the productivity of Copilot or Cursor without vendor lock-in or training-data exposure on proprietary code.

Why AI Code Generation IDE Plugin?

GitHub Copilot dominates with 42% market share but is trained on public code and performs weaker on private codebases without additional configuration; users also report confusing licensing and unexpected billing.
Cursor leads on revenue (USD 2B ARR) but requires switching away from JetBrains, a significant adoption barrier for enterprise Java/Kotlin teams, and offers no on-premises deployment.
Tabnine's Organisation Context Engine is gated behind enterprise pricing (USD 39–59/user/mo) and only fully realised after long indexing periods.
Amazon Q Developer announced end-of-support for its IDE plugins (retiring April 2027), creating procurement risk for AWS-aligned customers.
Continue is open source and model-agnostic but lacks polished UX, hosted model management, enterprise support, or compliance certifications, leaving a gap for a polished open alternative.

Key Features

Inline Completion and Chat

Ghost-text inline code completion (single-line and multi-line) for Python, JavaScript/TypeScript, Java, Go, C++, and Rust.
Context-aware chat panel scoped to the open file and current selection.
Code explanation and documentation generation.
Unit test generation from existing functions.

Codebase Context and Private-Repo Awareness

Codebase-wide context retrieval via RAG over a local repository index.
Organisation context engine that learns internal APIs, naming conventions, and architectural patterns from private repositories.
Configurable model backend supporting OpenAI, Anthropic, and local Ollama deployments.

Agentic Multi-File Execution

Multi-file task execution with diff-based approval UI before changes are applied.
MCP server protocol support for connecting external tools and data sources.
Step-by-step implementation plans surfaced for human review.

Security-First Generation

Inline security vulnerability flagging against OWASP Top 10 patterns as code is written, rather than only after generation.

IDE Coverage

Primary VS Code plugin surface with a JetBrains plugin as the secondary target.

AI-Native Advantage

The differentiation is AI applied where incumbents are weakest: organisation-specific fine-tuning that learns private APIs and conventions as a first-class capability rather than an enterprise add-on; agentic test generation and PR review that goes beyond completion; OWASP-aware inline security checks during coding instead of post-generation scans; and framework-agnostic legacy modernisation (e.g. Java 8 to 21, jQuery to React) that is not tied to a single cloud vendor.

Tech Stack & Deployment

The plugin targets VS Code and JetBrains IDEs and integrates with industry standards including the Language Server Protocol (LSP), Tree-sitter for AST access, the OpenAI fill-in-the-middle (FIM) completion format, the Model Context Protocol (MCP) for tool integration, and the GitHub Pull Request API for PR-level workflows. Configurable model backends allow OpenAI API, Anthropic API, or local inference via Ollama, supporting SaaS, self-hosted, and air-gapped deployment patterns.

Market Context

The AI coding assistant market reached USD 12.8 billion in 2026 and is projected to grow to USD 30.1 billion by 2032 at a 27% CAGR, with year-over-year growth of approximately 65% from 2025 to 2026. Per-seat pricing ranges from USD 10/user/mo (Copilot Individual) to USD 39/user/mo (Copilot Enterprise), with Tabnine on-prem at USD 39–59/user/mo. Primary buyers are individual developers on free tiers, engineering managers at growth-stage companies, enterprise CIOs standardising AI tooling, and security-conscious leaders requiring on-prem or private-cloud deployment.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.