Ix provides a developer tool that transforms a code repository into a structured, searchable graph, enabling instant understanding of codebases. Users run a single command to map source files, then query symbols, definitions, impacts, and history through a CLI, with persistent memory that retains context across sessions. The platform integrates with AI agents like Claude, Gemini, and Codex to enhance code exploration and issue detection.
Funding
Funding not disclosed
Founders
Product
Problem
Developers and AI agents spend significant time manually reconstructing a codebase’s architecture, dependencies, and impact of changes, often rereading source files for each task. This repetitive context building slows onboarding, increases risk of errors, and leads to inefficient use of computational resources.
Solution
Ix creates a persistent, graph‑based memory of a repository by running a one‑time `ix map` command that parses source files with Tree‑Sitter and builds a structured knowledge graph of symbols, relationships, and system boundaries. The resulting graph can be queried with concise CLI commands—such as `search`, `locate`, `explain`, `impact`, `trace`, and `smells`—to retrieve definitions, understand behavior, assess change blast radius, and detect architectural issues without rereading code. The graph is stored locally, keeping the code and its metadata on the developer’s machine, and remains available across sessions, enabling both humans and LLM‑based agents to maintain context and reduce token consumption. Ix integrates with popular AI agents (Claude, Codex, Cursor) via plugins and hooks, allowing agents to inject graph context into prompts, automatically update the graph after edits, and execute multi‑step workflows for impact analysis, refactoring planning, and documentation generation.
Target Audience
Ix is aimed at software engineers, DevOps teams, and AI‑assisted development tools that need rapid codebase understanding, safe change planning, and enhanced agent productivity.
Features
- One‑time `ix map` operation builds a living system graph from any supported language (14+ languages) using Tree‑Sitter parsing
- Persistent local memory retains the graph and session context across multiple CLI or agent interactions
- Rich query commands (`search`, `locate`, `explain`, `impact`, `trace`, `smells`, `rank`, etc.) provide fast, token‑efficient answers without source file reads
- Agent integrations via plugins and hooks for Claude, Codex, Cursor, and custom agents, enabling automatic context injection and graph updates during code edits
- Automated impact analysis and blast‑radius calculation to assess risk of changes before they are made
- Architecture health checks that detect code smells, coupling hotspots, and rank important components
- CLI and API access allow embedding Ix into existing developer workflows, CI pipelines, and custom tooling