Learning web development - Not a great "deal" - very few tokens and locked out future integrations below tier 4
Pros:
- Integrates knowledge bases and your repo giving context for the AI
- Right inside VS code, not only a chat, but with context
- Cool when you ask it to autocomplete in a small section of the code, but has missed multiple times (still cool though, writes the missing code for you!)
Cons:
- Much worse off deal than last deal, doesn't match the pricing plans on their website
- Don't know what model it is using
- Wrong answers
- Took over 2 days to log in
- Very low amount of tokens compared to other tools
- Tier 3 - for teams, 300K tokens isn't nearly enough, expect to use your own tokens
- Very expensive deal at $579 to get future integrations
- Low amount of knowledge bases (VERY low compared to last deal)
Overall, it's an OK software that will be developed (unlike other tools that came here that look more stagnated in terms of context, etc) but the amount of tokens is very low and the prices are sky high compared to past deals... Can work together with other tools in your toolset but the amount of tokens isn't anywhere close to make it a daily companion, nor the integrations, nor the amount of knowledge bases.
Ayush_CodeMate
Jun 15, 2024Hi @danirogerc
Thanks for the constructive feedback! We're always open to hearing from our users to improve our product experience. Answering your questions for better clarity-
- We understand our last deal comes with unlimited tokens, but we returned to AppSumo to come up with a deal which is sustainable for us in the longer run, since we're incurring high cloud costs running millions of queries every day, giving away unlimited tokens would not be sustainable for us in the longer run.
Hence to serve our users for lifetime, we came up with generous token limits every month, which we even increased yesterday to 750k in Tier 3, 3.75 million in Tier 4 and 20 million tokens in Tier 5, with an option to add 1 M tokens/month at just $60/yr, which is more than 80% less than GPT4-turbo and better performance.
- We use our own ML Models trained on proprietary datasets to maintain the desired accuracy and speed. We're also open sourcing some of our models to ensure transparency over our architecture. Checkout our first open source model on HF having over 2500 downloads every month-
https://huggingface.co/codemateai/CodeMate-v0.1
- Wrong answers- Can you share more information on the same (screenshot/video) at support@codemate.ai for us to look upon it closely?
- We have also added future integrations in our Tier 2 plan, the highest tier for individuals based on user requests.
- Low amount of knowledge bases (VERY low compared to last deal)- Apart from the token consumption, all other features are made better than our last deal. The knowledge base upload limits are the same as our last deal. Can you clarify once on this, what you're referring to?