See live cache hit and cost savings while running AI code review on your repo.
Compare cold versus warm cache performance across two passes on the same files.
Automatically re-review changed files every time you save while coding.
Review only the files changed in a pull request or against a target branch.
| diversexl/critcache | 0xradioac7iv/tempfs | 52191314/web-agent-proxy-sdk | |
|---|---|---|---|
| Stars | 0 | 0 | 0 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | easy | moderate | moderate |
| Complexity | 2/5 | 3/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Mock mode needs no account, the full experience requires a free BTL Runtime API key and Node.js 18 or newer.
critcache is a command line tool that reviews the code in your project by sending files to AI models through a service called BTL Runtime, and then shows you, live in your terminal, how much of that review was served from cache and how much money that cache saved you. The core idea is that BTL Runtime already tracks whether each AI response came from a cache or was freshly generated, along with the cost and time saved, but those details are normally hidden in technical response headers most people never check. critcache pulls that information out and displays it plainly as the tool runs. When you point it at a repository, it picks out the most relevant source files, sends each one to an AI reviewer at the same time rather than one after another, and prints a line per file showing whether it was a cache hit or miss, how long it took, and how much was saved. It also writes a full written report to a markdown file. A separate compare command runs the same analysis twice back to back on the same files, so you can see the cache hit rate and total time improve between the first, cold pass and the second, warm pass. Other commands let you watch a project continuously and re-review files automatically each time you save, review only the files that changed compared to another branch or commit, review your current uncommitted changes, check which AI providers connected to BTL Runtime are healthy, and get a score on how well your own prompt setup is likely to benefit from caching. You can try it without any account by running it in a mock mode that simulates the whole process locally with fake cache behavior. For the full experience, you need a free API key from the BTL Runtime service. It requires Node.js version 18 or newer, and there are no license details included in what was shown.
A CLI that runs parallel AI code reviews through BTL Runtime and shows you, live, how much caching is saving you in time and money.
Mainly TypeScript. The stack also includes TypeScript, Node.js, BTL Runtime.
No license information was included in what was shown, so terms of use are unclear.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.