Skip to content

Eden AI: simplify the use and deployment of AI technologies by providing a unique API that connects to the best possible AI engines

License

Notifications You must be signed in to change notification settings

edenai/edenai-apis

Repository files navigation

License Tests

Eden AI Logo

Table of Contents

Eden AI APIs

Eden AI aims to simplify the use and deployment of AI technologies by providing a unique API that connects to all the best AI engines.

With the rise of AI as a Service , a lot of companies provide off-the-shelf trained models that you can access directly through an API. These companies are either the tech giants (Google, Microsoft , Amazon) or other smaller, more specialized companies, and there are hundreds of them. Some of the most known are : DeepL (translation), OpenAI (text and image analysis), AssemblyAI (speech analysis).

There are hundreds of companies doing that. We're regrouping the best ones in one place !

➡️ Read more about it ...

why aren't you regrouping Open Source models (instead of proprietary APIs) into one repo? : Because it doesn't make sens to deploy and maintain large pytorch (or other framework) AI models in every solution that wants AI capabilities (especially for document parsing, image and video moderation or speech recognition) . So using APIs makes way more sens. Deployed OpenSource models are being included using different APIs like HuggingFace and other equivalents.

EdenAI Gif

Package Installation

You can install the package with pip :

pip install git+https://github.com/edenai/edenai-apis 

Quick Start

To make call to different AI providers, first add the api-keys/secrets for the provider you will use in edenai_apis.api_keys.<provider_name>_settings_templates.json, then rename the file to <provider_name>_settings.json

When it's done you can directly start using edenai_apis. Here is a quick example using Microsoft and IBM Keyword Extraction apis:

from edenai_apis import Text

keyword_extraction = Text.keyword_extraction("microsoft")
microsoft_res = keyword_extraction(language="en", text="as simple as that")

# Provider's response
print(microsoft_res.original_response)

# Standardized version of Provider's response
print(microsoft_res.standardized_response)

for item in microsoft_res.standardized_response.items:
    print(f"keyword: {item.keyword}, importance: {item.importance}")


# What if we want to try an other provider?
ibm_kw = Text.keyword_extraction("ibm")
ibm_res = ibm_kw(language="en", text="same api & unified inputs for all providers")


# `original_response` will obviously be different and you will have to check
# the doc of each individual providers to know how to parse them
print(ibm_res.original_response)

# We can however easily parse `standardized_response`
# the same way as we did for microsoft:
for item in ibm_res.standardized_response.items:
    print(f"keyword: {item.keyword}, importance: {item.importance}")

Asynchronous features

If you need to use features like speech to text, object extraction from videos, etc. Then you will have to use asynchronous operations. This means that you will first make a call to launch an asynchronous job, it will then return a job ID allowing you to make other calls to get the job status or response if the job is finished

from edenai_apis import Audio

provider = "google" # it could also be assamblyai, deepgram, microsoft ...etc
stt_launch = Audio.speech_to_text_async__launch_job(provider)
stt_get_result = Audio.speech_to_text_async__get_job_result(provider)


res = stt_launch(
    file=your_file.wav,
    language="en",
    speakers=2,
    profanity_filter=False,
)

job_id = stt_launch.provider_job_id

res = stt_get_result(provider_job_id=job_id)
print(res.status)  # "pending" | "succeeded" | "failed"

Available Features & Providers

⚠️ You can find a list of all available features and providers here ⚠️

Contribute

We would love to have your contribution. Please follow our guidelines for adding a new AI provider's API or a new AI feature. You can check the package structure for more details on how it is organized. We use GitHub issues for tracking requests and bugs. For broader discussions you can join our discord.

Don’t want to create accounts for all providers and host the project by yourself?

You can create an account on Eden AI and have access to all the AI technologies and providers directly through our API. Eden AI Logo

Join the community!

Join our friendly community to improve your skills, focus on the integration of AI engines, get help to use Eden AI API and much more !

Linkedin Medium

License

Apache License 2.0