Urgent.News

What's breaking now, across thousands of outlets.

Tech

Introducing Celmis: self-hosted code intelligence over a symbol graph

Celmis is self-hosted code intelligence. It reads a set of repositories once and keeps a symbol graph of them; asking questions, reviewing pull requests, auditing dependencies, routing alerts and serving an MCP endpoint are then all different ways of reading that one index, rather than five products each holding their own copy of your code. It runs on one machine under docker compose , with…

Celmis is a self-hosted code intelligence tool that creates a symbol graph of repositories and answers questions, reviews pull requests, audits dependencies, routes alerts, and serves an MCP endpoint using the same index. It runs on one machine using Docker Compose and can be powered by any model provider. Unlike diff-only tools, Celmis can answer questions spanning multiple repositories and identify matches across language boundaries.

It was introduced with an example of a diff-only tool that couldn't find the answer to a question that spanned two repositories, resulting in duplicated contract information in separate repositories. Celmis can help catch real issues, such as type errors and off-by-one errors, with inline suggestions and eliminates false positives.

Additionally, it can serve as an MCP server, interfacing with eighteen tools using the same symbol graph.

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 Tech

Whoosh on the no-GIL Python: pure-Python search that can finally use your cores

For fifteen years, the standard answer to "why is my pure-Python search slow to index?" was: the GIL . Tokenizing, stemming, and building postings is all CPU-bound Python, and only one thread gets to…

  • Python 3.14 introduced no-GIL build for pure-Python CPU-bound tasks
  • Whoosh search library benefits from no-GIL for parallel indexing
  • Concurrency contract requires one writer per thread to avoid lock contention

Code Is Open, Secrets Aren't: Organizing Your Project for AI Agents

This Article is also found in Web-Warrior-Toolbox Anything you put in the frontend or expose through routes is visible to users and AI agents, while secrets like API keys and credentials must stay…

  • Separate .env files for different environments to secure secrets
  • AI agents can access code repositories, not .env files
  • build.sh script builds project using appropriate .env file

More from Thursday 27 August →