|
<< Click to Display Table of Contents >> Navigation: 3. Script Language > AI - Artificial Intelligence Commands > AIG. - Google AI > 8. Results & Analysis > AIG. - AI Google Gemini Integration |
MiniRobotLanguage (MRL)
AIG.Get Tokens
Retrieve the Total Token Count
Intention
GetTokens Command: Measure API Usage
The GetTokens Command retrieves the total number of tokens (Input + Output) consumed by the most recent Google Gemini API transaction.
This value is extracted from the usageMetadata field returned by Google and stored in the internal AIG_TotalTokens variable.
It is crucial for monitoring costs and ensuring you stay within the model's context window limits.
The GetTokens Command simply reads the integer value representing the "Total Token Count" of the last operation.
This includes:
�Prompt Tokens: The cost of your question and any uploaded images/files.
�Candidate Tokens: The cost of the text generated by the AI.
Monitoring tokens is important for:
�Cost Control: Paid plans charge per 1 million tokens. This helps you calculate costs.
�Optimization: Identifying overly verbose prompts or responses.
�Logic Flow: Deciding if a conversation history is getting too large to send again.
Call the command immediately after AIG.Ask, AIG.AskVision, or AIG.GenerateImage.
You can provide a variable to store the count, or omit it to leave the value on the Top of Stack (TOS).
Example Usage
AIG.Ask|Write a short poem about coding.
AIG.GetTokens|$$TOK
AIG.GetContent|$$RES
DBP.Poem: $$RES
DBP.Cost in tokens: $$TOK
This script asks the AI a question, then retrieves both the answer and the "cost" (tokens used) to generate that answer.
Illustration
Prompt: "Hi" (1 token)
Reply: "Hello!" (2 tokens)
--------------------------
AIG.GetTokens -> 3
Syntax
AIG.GetTokens[|P1]
AIG.gtk[|P1]
Parameter Explanation
P1 - (Optional) Variable to store the numeric token count. If omitted, the value is placed on the Top of Stack (TOS).
Remarks
- Requires a prior API call. Returns 0 if no call has been made yet.
- Images count as a fixed token amount (approx. 258 tokens per image for Gemini 1.5 models).
Limitations
- Only reflects the usage of the single most recent request. It does not sum up session totals.
See also:
? AIG.Ask