fix the way to get API Key

replace SpeechToTextTool with AssemblyAIAudioTranscriptTool

Create assemblyai_audio__transcript_tool.py

improve custom prompt by specifying "Attached files can be downloaded via the Internet."

add additional_authorized_imports openpyxl

Specify CUSTOM_PROMPT to improve answer reproducibility

add additional_authorized_imports pandas

clarify additional_args

remove DuckDuckGoTool from tool list

set argument name explicitly in agent.run()

add transformers extra

import transforerms explicitly

add extra to avoid error

import extra seperately to avoid docstring related error

remove unnecessary import

add audio extra

allow agents to access files and add a tool to process audio files

remove custom_prompt because it does not appear to contribute to improved accuracy

remove planing_interval again add instructions

Update requirements.txt

remove use_structured_outputs_internally=True and restore planning_interval

remove planning_interval=3 because of bug and add use_structured_outputs_internally=True instead

set planning_interval =3

add import

stop hard-coding model_id

update default parameters

use smolagents's step_callbacks instead of litellm's callback

Rename api_call_counter.py to rate_limiter.py

use smolagents' step_callbacks instead of litellm's callback

remove unnecessary import

Rename rate_limiter.py to api_call_counter.py

fix variable name

wrap LiteLLMModel model to add rate limit logic

remove unnecessary "f" char

fixed variable name

reflect RateLimiter's changes

simplify counting and waiting logic

print before waiting

add some tools

add script for YoutubeTranscriptTool

add some libraries to add tools

remove unnecessary import

fixed module name typo

reflect code changes

rename APICallCounter to RateLimiter and refactor

add markdownify to use VisitWebpageTool

stop using add_base_tools=True and set tools explicitly

fix wrong method name

add logic to wait 60s after calling API 10 times

add reset method