HaiAI123

Curated Global AI Tools Directory

Industry update
Flat

15-year-old developer open-sources T-Zero AST pruning MCP server

11/21/2026 — 11/21, 08:18·1 sources (all)·1 reports (all)

Key takeaways

  • Turkish 15-year-old developer Toprak released T-Zero, an open-source local MCP server that prunes multilingual code down to typed AST interf…

Background & analysis

On November 21, 2026, at 07:42, the AI section of DEV Community published a post reporting that Toprak, a 15-year-old developer from Turkey, has open-sourced T-Zero, a local MCP server, with development work carried out at Siber Akademi. In his own account, Toprak says he regularly uses AI coding assistants such as Cursor, Claude Desktop, and Windsurf and repeatedly runs into a persistent obstacle on large codebases, which he describes as context window bloat.

His central argument separates implementation from interface. When a directory is referenced with the @ tag in Cursor Composer, he estimates that 85-90% of the tokens ingested are function execution bodies, covering loops, SQL queries, and helper transforms, none of which the AI actually needs to see. T-Zero responds by stripping multi-language code down to a typed AST interface skeleton, so that only interface-level information is passed to the model.

The published test results come from a repository containing 481 files. The original context of roughly 11 million tokens is reduced to 1.32 million, a drop of 88%, and per-file skeletons are cut by 94.6%. T-Zero is released as an open-source tool, and the report provides no information on future maintenance plans or version updates, so the project's current status rests at that initial open-source release.

AI-generated from 1 reports · updated 1 day ago

updream
24-hour heatHeat index 46 · peak 91 · 23h ago
24 hours agonow

Reports on this story headlines open the original

Yesterday
  1. Toprak, a 15-year-old developer in Turkey, has open-sourced T-Zero, a local MCP server that prunes multi-language code into typed AST interface skeletons. On a 481-file repository, it reduced raw context from 10,998,573 tokens to 1,319,828, an 88% cut, with a focus-file skeleton reaching 94.6% reduction. The tool targets context-window bloat in assistants such as Cursor and Windsurf.

    DEV Community · AIAI score 60

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →