Skip to main content

A mini framework built on top of Langchain and Llamaindex to provide LLM powered Autonomous Agents as a simplified service to assist users with their tasks

Project description

ServifAI

A mini framework built on top of Langchain and Llamaindex to provide LLM powered Autonomous Agents as a simplified service to assist users with their tasks.

PyPI PyPI

Overview

ServifAI (Task-based Agent) = LLM + Memory + Planning + Toolbox

agent_pic

Instead of feeding all kinds of tools to a single agent and confusing it while selection, ServifAI narrows down the selection by combining only necessary tools on basis of the task at hand.

Read this article to get an overview on Agents.

Currently we only support OpenAI models. If you are privacy concerned you should apply for Azure OpenAI services. The reason for not yet supporting opensource local models are:

  • You need a lot of money - either to buy a great GPU or host it on cloud with a great GPU
  • even if you tick the first point, it won't help you much as currently opensource models are good for chat but lag behind OpenAI models in enabling complex agent workflows

Current Supported Tasks:

Tasks Toolbox Tools Required File Extensions
default DuckDuckGo + LLM Math + PAL Math None
qna_local_docs Vector Index + Knowledge Graphs PDF/DOCX

Installation

Works best with Poetry

poetry add servifai

With pip, you might have to install dependencies manually

pip install servifai

Usage

Create a .env file for openai

OPENAI_API_KEY='sk-...'

Run Python Code

from servifai import ServifAI
myagent = ServifAI()

while True:
    text_input = input("Me: ")
    if text_input == "exit":
        break
    response = myagent.chat(text_input)
    print(f'ServifAI: {response}\n')

Output

Me: I am feeling bored at home, provide me a list of places in bengaluru to chill out on a sunday.
ServifAI: Here is a list of places in Bengaluru to chill out on a Sunday:

1. Cubbon Park
2. Rasta Cafe
3. Skandagiri
4. Uttari Betta Sunrise
5. Ranganathaswamy Betta
6. Adventure Camping at Nandi Hills
7. Riverside Manchanabele
8. Caving at AntaraGange
9. Clubbing at Hard Rock Café
10. High Ultra lounge
11. Toit
12. Dinning at Empire Hotel

These are just a few options, and there are many more places in Bengaluru where you can relax and have a good time on a Sunday.

Me: Too many options for a day. Provide a itinerary instead for one day
ServifAI: Based on the search results, here is a suggested one-day itinerary for Bengaluru:

Morning: Start your day with a visit to Cubbon Park and enjoy the greenery. Join the locals on their early morning walk.

Afternoon: Indulge in a classic South Indian breakfast and then head to Savandurga, a famous place known for its temples and rock climbing.

Evening: Explore the vibrant streets of Bengaluru and visit popular spots like MG Road or Brigade Road for shopping and dining.

Night: End your day by experiencing the nightlife of Bengaluru at one of the city's popular clubs or lounges.

Please note that this is just a suggested itinerary and you can customize it based on your preferences and interests.

Me: Do you think this would be possible to complete based on today's weather?
ServifAI: Based on the weather forecast, it is possible to complete the activity you mentioned. The maximum temperature in Bengaluru is expected to be around 32 degrees Celsius today.

Me: exit

Check Examples for more.

TODO:

  • Add support for popular unstructured data formats
  • Add support for other VectorDBs
  • Add other task based tools
  • Add support for structured data formats
  • Support for OpenAI funcs

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

servifai-0.2.3.tar.gz (10.7 kB view details)

Uploaded Source

Built Distribution

servifai-0.2.3-py3-none-any.whl (11.9 kB view details)

Uploaded Python 3

File details

Details for the file servifai-0.2.3.tar.gz.

File metadata

  • Download URL: servifai-0.2.3.tar.gz
  • Upload date:
  • Size: 10.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.5.1 CPython/3.10.8 Linux/5.15.0-1041-azure

File hashes

Hashes for servifai-0.2.3.tar.gz
Algorithm Hash digest
SHA256 dbc2133cbde0e163c188ad2c9a52bf8a8083507951125ba10f51b3eaaa846ebf
MD5 29956669f633b3c0e73372434975a3c8
BLAKE2b-256 a9fcaad1e356d9af5b5d186670e779643d56af360a5f5404cc85734f492baadd

See more details on using hashes here.

File details

Details for the file servifai-0.2.3-py3-none-any.whl.

File metadata

  • Download URL: servifai-0.2.3-py3-none-any.whl
  • Upload date:
  • Size: 11.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.5.1 CPython/3.10.8 Linux/5.15.0-1041-azure

File hashes

Hashes for servifai-0.2.3-py3-none-any.whl
Algorithm Hash digest
SHA256 0f81e8be121461a64ecc8ffa99d46a0da99b64518d227f05ef46ea157a738497
MD5 cafbb684bd4148949220f1e0fd425822
BLAKE2b-256 630ab0c325b5a8abb0d408f040465526b703f45ba6a6f8a2a6738468174d0f77

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page