<< Click to Display Table of Contents >> Navigation:  »No topics above this level« AIN. - AnythingLLM AI

---

MiniRobotLanguage (MRL)

 

AIN.SetMaxToken
Set the Maximum Token Limit for AnythingLLM API Calls

 

Intention

 

SetMaxToken Command: Controlling Response Length
 
The SetMaxToken command allows you to define the maximum number of tokens (words or word pieces) that the AnythingLLM API can generate in a response.

This helps manage response length and resource usage in your scripts.

It’s part of the AIN - AnythingLLM AI suite.

 

What is the SetMaxToken Command?

 

This command updates the AIN_MaxToken global variable, which caps the token count for API responses.

If set to 0 or omitted, it defaults to 4096 tokens, as defined by %AIN_MaxToken.

 

Why Do You Need It?

 

Setting a maximum token limit is useful for:

•Efficiency: Limit response size to reduce processing time and resource consumption.

•Relevance: Prevent overly long responses that may dilute key information.

•Control: Tailor output length to fit specific script requirements.

 

How to Use the SetMaxToken Command?

 

Use this command before making API calls like AIN.Ask or AIN.AskV to control response length.

The value must be a positive integer; otherwise, it defaults to 4096.

 

Example Usage

 

AIN.SetMaxToken|100

AIN.Ask|Explain quantum physics|$$EXP

DBP.Short Explanation: $$EXP

 

This limits the response to 100 tokens for a concise explanation.

 

Illustration

 

┌────────────────────┐

│ Max Tokens         │

├──────┬─────────────┤

│ 100  │ Short       │

│ 4096 │ Default     │

└──────┴─────────────┘

Setting max tokens to control response length.

 

Syntax

 

AIN.SetMaxToken|P1

AIN.Set_MaxToken|P1

 

Parameter Explanation

 

P1 - The maximum token limit (positive integer). If 0 or omitted, defaults to 4096.

 

Example

 

AIN.SetMaxToken|500

AIN.Ask|Summarize a book|$$SUM

DBP.Book Summary: $$SUM

ENR.

 

Remarks

 

- Verify the current limit with AIN.GetMaxToken.

- The actual token count may vary slightly depending on the model’s tokenization.

 

Limitations

 

- The maximum value is model-dependent; exceeding the model’s limit may truncate responses.

- Negative values are not supported and reset to the default (4096).

 

See also:

 

• AIN.GetMaxToken

• AIN.Ask

• AIN.AskV

• Max Token Configuration