AnonniX / phind-AI-API

Unofficial Phind API documentation

Unofficial Phind API Documentation

v1.0.1, last updated 15/05/2023 18:22 BST

Disclaimer

This documentation is for educational purposes only. It may be incorrect, outdated, or be missing features. Some aspects are unconfirmed, such as the request option parameters.

Got any feedback or want to add to the docs?

That's great! Please create a pull request. Thank you :)

What is Phind?

Phind is an AI service that provides natural language responses to user questions and prompts. It has a low-powered model that rewrites user prompts and then uses search results to generate an answer with GPT-4, the higher powered model used when 'Use Best Model' is checked. This documentation does not cover intermediate mode, which is presumed to use gpt-3.5-turbo or text-davinci-003.

Base URL

http://phind.com/api

API Endpoints

/api/web/search

Search the web based on a user's question or prompt.

Method: POST

Request Body: JSON

{
  "q": "String query",
  "browserLanguage": "String language code"
}

Response: JSON

[
  {
    "title": "String title",
    "url": "String url",
    "is_source_local": Boolean,
    "is_source_both": Boolean,
    "description": "String description",
    "profile": {
      "name": "String name",
      "url": "String url",
      "long_name": "String long name",
      "img": "String image url"
    },
    "language": "String language code",
    "family_friendly": Boolean,
    "type": "String type",
    "subtype": "String subtype",
    "meta_url": {
      "scheme": "String scheme",
      "netloc": "String netloc",
      "hostname": "String hostname",
      "favicon": "String favicon url",
      "path": "String path"
    },
    "thumbnail": {
      "src": "String thumbnail url",
      "original": "String original thumbnail url",
      "logo": Boolean
    },
    "age": "String published date",
    "article": {
      "author": [
        {
          "type": "String type",
          "name": "String name",
          "url": "String url"
        }
      ],
      "date": "String published date",
      "publisher": {
        "type": "String type",
        "name": "String name",
        "url": "String url",
        "thumbnail": {
          "src": "String publisher logo url",
          "original": "String original publisher logo url"
        }
      }
    }
  }
]

/api/db/cache

Cache data in the Phind database. Used internally.

Method: POST

Request Body: JSON

{
  "key": "String key",
  "value": "String value"
}

Response: No response

/api/infer/answer

Generate an answer from the Phind AI model using web search results.

Method: POST

Request Body: JSON

{
  "question": "String question",
  "webResults": [web search response],
  "options": {
    "temperature": Float,
    "max_tokens": Integer,
    "top_p": Float,
    "repetition_penalty": Float,
    "presence_penalty": Float,
    "frequency_penalty": Float,
    "curiosity_bonus": Float
  }
}

Response:

data: String generated response, by word (10 word output = 10x data: word lines)

/api/infer/followup/answer

Generate a follow-up answer from the Phind AI model using web search results and conversation context.

Method: POST

Request Body: JSON

{
  "question": "String question",
  "questionHistory": ["String previous questions"],
  "answerHistory": ["String previous answers"],
  "webResults": [web search response],
  "options": {
    "temperature": Float,
    "max_tokens": Integer,
    "top_p": Float,
    "repetition_penalty": Float,
    "presence_penalty": Float,
    "frequency_penalty": Float,
    "curiosity_bonus": Float
  }
}

Response:

Same as /api/infer/answer