Urgent.News

What's breaking now, across thousands of outlets.

AI

Your AI Coding Agent Is Blind: Gortex Gives It a Map of Your Codebase

Your AI coding agent reads your codebase like a tourist reads a city: one street at a time, no map, asking for directions after every turn. Every task starts with the same ritual. The agent runs a grep, opens a file, reads 400 lines to find one function, follows an import, opens another file. By the time it writes a line of code, half of your context window is gone, and you are paying for every…

AI coding agents often struggle with understanding the broader context of your codebase. They read files one at a time, often requiring extensive token usage before they can write any code. This inefficiency can lead to costly token consumption. A new open-source project called Gortex aims to address this issue by indexing your codebase into a persistent knowledge graph and providing it to AI agents through MCP (Multi-Component Protocol).

Gortex parses your code using tree-sitter and stores the results as a graph, which includes information about functions, classes, call chains, HTTP routes, and cross-service contracts. The tool supports 257 languages with hand-tuned tree-sitter queries for the most popular ones. It exposes the graph through an MCP server with 175 configurable tools that allow the agent to perform various tasks such as finding usages, get callers, or blast-radius queries.

One of the most interesting features of Gortex is its ability to detect API contracts across repositories. It recognizes different frameworks and annotations, allowing it to normalize and match contracts to identify orphan providers, consumers, and mismatches. This feature is particularly useful for developers working in a microservice environment.

The project claims up to 50 times fewer tokens per response, which can lead to significant cost savings compared to traditional AI coding agents. The benchmarks from the project measure the indexing time, memory usage, and token savings for various codebases, demonstrating the effectiveness of the graph-based approach. The installation process for Gortex is straightforward, with a single command to set it up on your machine.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

A Conversation ID Is Not Project Context

A Conversation ID Is Not Project Context A conversation ID can feel like the fastest way to give an agent continuity: save it in the repository, pass it to the next machine, and resume where the last…

Best AI Research Tools 2026: What Works

Everyone has a "deep research" button now. ChatGPT has one. Gemini has one. Perplexity built its whole identity on it. And here is the thing nobody tells you at the door: one button is not enough.

  • ChatGPT excels in cautious responses and citation accuracy, but can be slow
  • Gemini offers broad coverage for Google ecosystem users, but generates unreliable links
  • Perplexity performs best in citation accuracy with 37% failure rate, excels in speed

More from Sunday 27 September →