based
088814edd6
also reverted the azure checker to being multithreaded instead of asynchronous, the async version does nothing but choke on large key counts even with a small semaphore value. if theres any sort of rate limit in place on something then async is slop. run_in_executor was also causing timeout errors and ignoring caught exceptions so it's been removed and replaced with the old as_completed method. it gets passed to an executor in the end either way so there shouldn't be much of a performance difference, if any.
keychecker
a fast, bulk key checker for various AI services
Currently supports and validates keys for the services below, and checks for the listed attributes a key might have:
- OpenAI - (Best model, key in quota, RPM (catches increase requests), tier, list of organizations if applicable, trial key status)
- Anthropic - (Pozzed check)
- AI21 - (Trial check)
- Google MakerSuite (Gemini)
- AWS - (Admin status, auto-fetch the region, logging status, username, bedrock status)
- Azure - (Auto-fetch all deployments, auto-fetch best deployment/model, filter status)
- Google Cloud Vertex AI - (Requires a key file since oauth tokens expire hourly. Good luck scraping for those.)
- MistralAI - (Subscription status)
Always open to adding more services, although I think I've covered all the ones people care about. Nonetheless, feel free to open an issue or PR if you want something else added.
Usage:
pip install -r requirements.txt
python main.py
Optional Arguments:
-proxyoutput
Outputs keys in a format that can be easily copy pasted into khanon's proxy instead of pretty print
-nooutput
Stops outputting and saving keys to the snapshot file (proxyoutput will also do this)
-file
Reads keys from a file instead of stdin, place either the absolute or relative path to the file in quotes after the flag.
-verbose
Displays an output as keys are being checked real time.