API Documentation

Overview

Swiftask is an AI aggregator platform allowing users to access AI functionalities provided by multiple vendors through a unified interface. This document describes how developers can interact with the Swiftask API to send and receive AI-generated responses.

Authentication and Authorization

To ensure secure access to the Swiftask API, authentication and authorization are required for all API requests. Below are the steps to obtain and use an API key, which is necessary for accessing the AI Bot endpoint.

Obtaining an API Key

1. Create an Account: If you haven't already, you need to sign up for an account at Swiftask.

2. Generate API Key: Once logged in, navigate to the Developers space section of your account, there, you will find an option to generate a new API key. Click on this option and label your new key for easy identification later.

Developer Space

1. Record Your API Key: After generation, your API key will be displayed. It is crucial to copy and securely store this key as it will not be visible again. If you lose your API key, you will need to generate a new one.

Base URL

https://graphql.swiftask.ai/api

Endpoint Description

POST /ai/:slug

This endpoint is used to interact with a specific AI identified by a unique slug. It accepts JSON input containing session details and user input, and returns AI-generated responses.

URL Parameter:

• slug: Unique string identifier for the AI bot.

HTTP Method: POST

Headers:

• Content-Type: application/json

• Authorization: Bearer <swiftask_api_key>

Request Body:

{

  "input": "<user_input_message>",
  "documentAnalysisMode": "ADVANCED | SIMPLE",
  "files": [
    {
      "url": "<file_url>",
      "type": "<file_type>",
      "size": <file_size>,
      "name": "<file_name>",
      "id": <file_id>
    }
  ],
  "messageHistory": [
    {
      "role": "user|assistant|system",
      "content": "<message_content>",
      "files": [
        {
          "url": "<file_url>",
          "type": "<file_type>",
          "size": <file_size>,
          "name": "<file_name>",
          "id": <file_id>
        }
      ]
    }
  ]
}

Response:

{
  "text": "<response_text>",
  "botSlug": "<bot_identifier>",
  "files": [
    {
      "url": "<file_url>",
      "b64_json": "<base64_encoded_json>",
      "type": "<file_type>",
      "size": <file_size>,
      "name": "<file_name>",
      "id": <file_id>
    }
  ],
  "additionalMessages": [
    {
      "text": "<additional_message_text>"
    }
  ],
  "isBotError": <boolean>,
  "sessionId": <session_identifier>,
  "totalBotUsage": <total_usage_count>,
  "usageDetail": {
    "totalInputTokens": <number_of_input_tokens>,
    "totalOutputTokens": <number_of_output_tokens>
  },
  "isSubscriptionUsageLimitReachedMessage": <boolean>,
  "sourceMetadatas": [
    {
      "url": "<source_url>",
      "name": "<source_name>",
      "textChunk": "<text_chunk>",
      "dataSourceId": <data_source_id>,
      "type": "<source_type>",
      "page": "<source_page>"
    }
  ]
}

Sample curl request

curl -X POST 'https://swiftask.com/api/ai/{slug}' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer <api_key>' \
-d '{
  "input": "Hello, how can I help you today?",
  "files": [],
  "messageHistory": [
    {
      "role": "user",
      "content": "Hello, how can I help you today?"
    }
  ]
}'

Error Handling

Responses from this API will also include standard HTTP status codes to indicate success or failure of the requested operations. Common errors include:

• 400 Bad Request: Input data is malformed or validation failed.

• 401 Unauthorized: Missing or invalid authentication token.

• 404 Not Found: The specified AI slug was not found.

• 500 Internal Server Error: An error occurred on the server which was not the client's fault.

Usage Guidelines

• Ensure the session ID is unique for each session to avoid data conflicts.

• Handle file data carefully, especially when dealing with personal or sensitive information.

• Monitor your usage to avoid hitting subscription limits unexpectedly.

This API documentation covers the main aspects needed to integrate with various AI bots via Swiftask. For further details or support, contact the Swiftask support team.

Available AI

Below is a table listing each AI name along with its corresponding slug. This reference is crucial when making API requests to interact with different AI functionalities.

AI Name	Description	slug	API cost
Audio AI Transcription	Record or upload your audio file and get it transcribed, summarized, and translated.	audioia	3000.0	Minutes
ChatOnPDF	Interact with documents through conversation. Receive immediate responses complete with cited sources. Explore Documents in an unprecedented way with Swiftask.	chatpdf	5000.0	Questions
Claude 2.1	Claude 2.1 is the latest AI assistant model developed by Anthropic. It offers significant upgrades and improvements compared to previous versions.	claudev21	7.0	Words
Claude 3.5 Sonnet	Anthropic's latest AI model	claude35sonnet	6.0	Words
Claude 3 Haiku	Anthropic's Claude 3 Haiku: Outperforms models in its class for performance, speed, and cost without specialized fine-tuning.	claude3haiku	2.0	Words
Claude 3 Opus	Claude 3 Opus: Cutting-edge AI model with a 200K token context window. Unmatched performance and near-human comprehension for complex tasks.	claude3opus	25.0	Words
Cohere	Chatbot based on cohere model that can answer questions like ChatGPT	cohere	1.0	Words
DALL-E 3	DALL·E 3 is an AI model developed by OpenAI, which can generate highly realistic and detailed images from textual descriptions.	dallee3	31250.0	Images
Data Analysis Assistant	Data Analysis Assistant powered by OpenAI GPT-4o, to help you clean, analyze, and visualize your data. It can also create files, analyze images, and more	data-analysist	10000.0	Questions
DEMUCS	Demucs is a model that allows you to separate a music track into different components: bass, drums, vocals, guitar, and piano.	demucs	2000.0	Audio
DocumentAI	An AI agent specialized in extracting tables from files. Performs optical character recognition (OCR) and extracts data tables from PDF, PNG, JPEG files and other common formats.	document-ai	12500.0	Pages
Document Analyzer	Analyse, extract, summarize and generate insights from documents	invoice_analyzer	7.0	Words
ElevenLabs	Create the most realistic speech with AI	elevenlabs	40.0	Characters
Face Restoration	The Face Restoration Bot is a highly practical tool equipped with advanced algorithms designed to restore and enhance faces in old photos or AI-generated images.	facerestoration	5000.0	Images
Face To Many	Face to Many is a model that allows you to transform a face into various styles: 3D, emoji, pixel art, video game, claymation, or toy.	facetomany	5000.0	Images
Flux Pro	FluxPro is a model for image generation with top of the line prompt following, visual quality, image detail and output diversity.	fluxpro	22000.0	Images
GDocs	GDocs is a utility that helps you save chat text in a Google Doc, or create a new Google Doc, Google Sheet, or Google Slides presentation from natural language instructions.	gdrive	1.0	Export
Gemini Pro 1.5	Gemini Pro 1.5 is the next-generation model that delivers enhanced performance with a breakthrough in long-context understanding across modalities.	geminipro	3.0	Words
GPT-3.5	GPT-3.5: OpenAI's advanced language model, capable of intelligently understanding and generating text for various applications.	gpt35	1.0	Words
GPT-4o	OpenAI's multimodal model, fast, cost-effective, with excellent vision and multilingual performance.	gpt-4o	2.0	Words
GPT-4o mini	GPT-4o mini is the most advanced and cost-effective LLM	gpt-4o-mini	0.5	Words
GPT-4 Turbo	GPT-4 Turbo is more capable and has knowledge of world events up to April 2023. It has a 128k context window so it can fit the equivalent of more than 300 pages of text in a single prompt.	gpt4	10.0	Words
GPT4 Vision Turbo	GPT-4 Vision (GPT-4V) is a multimodal model developed by OpenAI. It allows the model to interpret and analyze images, not just text prompts, making it a "multimodal" large language model.	gpt4vision	10.0	Words
GPT Pro	GPT Pro is a general-purpose chatbot based on OPEN AI GPT model that can be used to chat on a variaty of documents files, and customised to your needs. It has access to Code-Interpreter	gpt4pro	6000.0	Questions
Live Portrait	Live Portrait is a model that allows you to animate a portrait using a driving video source.	liveportrait	21000.0	Images
Llama 3	Llama 3 is an open-source large language model (LLM) developed by Meta. It is designed for creating generative AI applications, including chatbots that can engage in natural language conversations and respond to a wide range of queries.	llama3	1.0	Words
MagicColor	Magic Color lets you colorize black and white images using AI	magiccolor	12500.0	Images
Meeting Transcription	Record or upload your audio file and get it transcribed, summarized, and translated.	audioainew	3000.0	Minutes
Meta Llama 3.1 405b	Llama 3.1 is a powerful, open-source AI model that can understand and generate human-like text in multiple languages, enhancing various applications.	llama31	6.0	Words
Mistral Codestral	Codestral is a cutting-edge generative model that has been specifically designed and optimized for code generation tasks, including fill-in-the-middle and code completion. Codestral was trained on 80+ programming languages, enabling it to perform well on both common and less common languages.	codestral	1.0	Words
Mistral Codestral Mamba	Codestral Mamba, a Mamba2 language model specialised in code generation	codestral-mamba	1.0	Words
Mistral Large	Mistral Large is introduced as the flagship language model by Mistral, boasting unrivaled reasoning capabilities.	mistrallarge	7.0	Questions
Mistral Medium	Mistral Medium is a versatile language model by Mistral, designed to handle a wide range of tasks. It features a 16K tokens context window and is natively fluent in multiple languages including English, French, Spanish, German, and Italian.	mistralmedium	5.0	Questions
Mistral Nemo	Mistral Nemo is an open source multilingual language model by Mistral, released in July 2024.	open-mistral-nemo	0.5	Words
o1	The OpenAI o1 model is an advanced AI that excels at solving complex problems. It thinks carefully before answering, using broad knowledge to reason through challenges in areas like math, science, and programming.	o1	10.0	Words
OCR	OCR allows extracting text from scanned images, PDFs or handwritten documents, and you can then interact with the extracted text. To get started, please upload the image or document you want to extract text from.	ocr	5000.0	Pages
OpenAI o1-mini	The OpenAI o1 AI model is designed to enhance reasoning capabilities by spending more time processing inputs before responding. It excels in complex tasks like science, math, and coding, performing at a level comparable to PhD students in benchmark tasks.	o1-mini	2.0	Words
OpenAI o1-preview	The OpenAI o1 AI model is designed to enhance reasoning capabilities by spending more time processing inputs before responding.	o1-preview	10.0	Words
Perplexity	Perplexity is an AI-powered search engine and conversational AI tool that aims to unlock the power of knowledge through information discovery.	perplexityonline	2500.0	Questions
PuLID	PuLID is an AI model that customizes images effortlessly while preserving their core features.	pulid	6500.0	Images
Runway Video Generator	Video Generator is a image to video model and can be directed with user prompt	runway	20900.0	Seconds
Scrapio	Scrapio is a chatbot that scrapes text from one or more web pages links that you provide. Talk to it in natural language to automatically extract the text contents you need.	scrapio	10000.0	Links
SQL Agent (BETA)	SQL Agent: Quickly obtain actionable insights from your database by simply asking questions.	sqlagent	1000.0	Questions
Stable Diffusion	The Stable Diffusion Bot is an innovative AI-powered tool that uses a text-to-image generative model to create stunning images from textual descriptions.	stablediffusion	10000.0	Images
Swiftask	General-purpose AI assistant bot powered by GPT-4o of OpenAI ChatGPT.	swiftask	0.001	Words
Text to Speech	Convert text to human-like speech	text2speech	1.0	Query
Text to video with Luma AI	Luma AI is a cutting-edge model that enable creating realistic 3D video directed with user prompt	lumaai	170000.0	Videos
Web Search	GPT based autonomous agent that does online comprehensive research on any given topic	webreporter	10000.0	Searches