# yc-duan/fastctx

Fast, context-efficient repository tools for AI agents (MCP)

Repository: https://github.com/yc-duan/fastctx
Canonical: https://ross.abutalabs.com/products/fastctx
Language: Rust
License: Apache-2.0
License Family: permissive
Last push: 2026-08-23T20:27:00+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 3
- inputs: {"age_days": 47, "days_push": 10, "days_rel": 10, "gap_med": 2.5, "n_releases_24m": 9}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1203, forks 50 (observed 2026-08-28T04:03:58.589505+00:00)

## What it is
FastCtx is a Rust-based local runtime that exposes fast, context-efficient repository tools (file reading, grep, glob, batch replacement, shell execution) to AI coding agents via MCP. It runs a persistent shared control center behind thin stdio proxies, reducing tool-call overhead and context consumption for agents like Codex.

## Use cases
- give coding agents fast file search and read tools via mcp
- reduce context window usage when agents explore a repository
- replace ad-hoc shell commands with structured tool calls for ai agents
- connect repository tools to codex cli or chatgpt
- batch replace code across a repo from an ai agent

## When to choose
- you use MCP-based coding agents and want fewer, cheaper tool calls
- you want stable structured inputs/outputs for repo operations instead of shell escaping
- you need reliable sessions with cancellation and reconnect behavior

## When to avoid
- you need a general-purpose IDE or standalone code search tool without an AI agent
- you only use agents without MCP support and cannot run fastctx serve

## Facets
- artifact type: cli-tool
- maturity: active
- function: mcp, search-engine, developer-tools, cli, file-system, parser
- domain: developer-tools, large-language-models
- platform: rust, cli, cross-platform
- tags: mcp-server, coding-agents, context-efficiency, repository-tools, codex, file-search, command-line, ai-agents, nodejs

## Member repositories
- yc-duan/fastctx (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:58.589505+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:20:33.504368+00:00, confidence not recorded.
  - readme: https://github.com/yc-duan/fastctx (fetched 2026-08-28T04:03:58.589505+00:00, sha f32b0fc07143)
- Data as of 2026-08-30T08:39:29.467469+00:00.