HN Showcase
Back to browse
Ctxsync – Chat with your codebase that stays in sync

Ctxsync – Chat with your codebase that stays in sync

by jelvibe25·Feb 13, 2026·2 points·0 comments
Visit ProjectView on HN

AI Analysis

●●SolidSolve My ProblemSlickShip It
The Take

Keeping an LLM's knowledge fresh across changing repos is a small but painful gap — ctxsync attacks it with scheduled and on-demand syncs and per-chat containers so sessions don't cross-contaminate. It surfaces exact file paths and line numbers, supports GitHub OAuth and BYO LLM keys, but I'd want more detail on embeddings, vector store choices, and access controls before trusting it on sensitive code.

Target Audience

Backend/frontend developers, engineering teams, and maintainers who want searchable, up-to-date LLM access to their code and docs

Post Description

Hi HN. I got tired of uploading my data to ChatGPT, having a useful conversation, then realizing a week later I need to re-upload again, because data changed.

Built ctxsync to fix it:

• Connect GitHub repos, documentation sites, files

• Sync on demand or schedule – your AI's knowledge stays current

• Data isolation: each chat runs in separate containers, no cross-contamination

Uses your own LLM API keys (OpenAI, Anthropic, Kimi-Code).

Early access: free

What do you think?

Similar Projects