Credits are used as follows:

• AI Model Input Tokens Costs:
  • 4 credits for 100k tokens using GPT-4O Mini.
  • 50 credits for 100k tokens using GPT-4O.
  • 5 credits for 100k tokens using Anthropic Claude Haiku.
  • 60 credits for 100k tokens using Anthropic Claude Sonnet 3.5.
  • (and so on…)
• AI Model Output Tokens Costs:
  • 15 credits for 100k tokens using GPT-4O Mini.
  • 200 credits for 100k tokens using GPT-4O.
  • 25 credits for 100k tokens using Anthropic Claude Haiku.
  • 300 credits for 100k tokens using Anthropic Claude Sonnet 3.5.
  • (and so on…)
• Video/Audio Training Costs:
  • 1 credit per minute for training with video (including YouTube) or audio files.
• Tool Invocation Costs:
  • Each tool has its own number of credits charged. We strive to keep the cost as low as possible so you get the most value at the right price. For more details please use our calculator above
  • Charges apply only when tools are used; no charge for unused configured tools.
• Training costs:
  • We charge 4 credits per 100 chunks monthly of text from your training data.
  • Example: A 2-page PDF typically contains 3 chunks, so training with 15 such PDFs (45 chunks) costs ~2 credits / monthly.

Of course not! As long as you have a subscription active, your credits will remain in your balance so every month you unblock the number of credits associated with your subscription.

For example, if at the end of the first month you have in your balance 1130 credits and you are on the Starter plan, in the next month you will receive 1200 more credits so you’ll have in your balance 2330 credits.

Credits from the Free Plan expire every month.

Your credits are topped up at the start of every calendar month regardless of when you subscribed. So if you subscribe on the 15th of June, your credits will be added to your balance on the 1st of July.

A token is a unit of text used by a language model (LLM), representing a piece of input or output. Tokens can be as small as a single character or as large as a word, depending on the context.

Example:

The sentence “AI is amazing!” consists of 4 tokens: “AI”, “is”, “amazing”, and “!”.

Not at all! Feel free to upload as many documents or videos as your heart desires.

That said, for optimal results, we suggest setting up specialized AI Assistants for different domains. Then, link all of them together in an AI Team. By dividing the workload among specialized assistants, you’ll enjoy more precise and accurate responses. It’s all about working smarter!

Absolutely! Uploading a YouTube video is super easy. All you need to do is pop in the video URL, and your AI Assitant will learn directly from that video!

But wait, there’s more!

When the AI Assistant uses your YouTube video to respond, you can watch the video right on our platform, or via a shareable link if you prefer. It’ll even jump straight to the exact minute the response is related to. Isn’t that convenient?

Absolutely! With our user-friendly no-code interface, you can seamlessly integrate your AI Assistant with your own API. This lets assistant perform actions or fetch data directly from your systems, making things super smooth for you.

Worried about data access? No problem! You can fine-tune what your assistant can see by setting up an extractor for your API response. For instance, if you have an API method GET /orders and you only want assistant to access the order’s status, you can simply extract the status. This way, only the necessary information is relayed back to the user.

Easy, secure, and efficient!