Augment Code
AI software development platform with a semantic Context Engine that indexes entire codebases — code, dependencies, architecture, and commit history — to provide deep context to AI models. Available as IDE extensions, CLI, AI code review, and as an MCP server for third-party agents.
Augment Code is built around its Context Engine, a semantic search system purpose-built for code that parses codebases to build dependency graphs, analyzes commit history, and uses embeddings to capture relationships beyond syntax. Rather than simple keyword or static analysis, the Context Engine indexes code, dependencies, architecture patterns, commit history, external documentation, and tickets to surface precisely what is relevant to a given task.
The product suite includes IDE agents for VS Code and JetBrains with multi-step task lists, a terminal-based CLI tool (Auggie), an AI code reviewer that posts inline GitHub comments, and Intent — a workspace for coordinated agents with living specifications and isolated environments. It supports JavaScript, TypeScript, Python, Go, and other major languages.
In February 2026, Augment exposed its Context Engine as an MCP server, enabling any MCP-compatible agent — including Claude Code, Cursor, and OpenAI Codex — to access its semantic indexing. Benchmarks claim a 71% improvement for Cursor and 80% for Claude Code when paired with the Context Engine. The MCP server runs locally via the Auggie CLI or remotely via Augment's hosted endpoint.
Founded by Igor Ostrovsky (ex-Pure Storage chief architect) and Guy Gur-Ari (ex-Google AI researcher), the company has raised $252M at a $977M valuation. It holds SOC 2 Type II and ISO/IEC 42001 certifications, never trains on customer code, and counts MongoDB, Spotify, Pure Storage, and Snyk among its customers. Pricing starts at $20/month (Indie, 125 messages) and $50/month (Developer, 600 messages), with custom enterprise plans available.