dsdanielpark/Bard-APIArchived
Bard API
Bard-API ist ein asynchroner Python-Wrapper und Client für die Interaktion mit Google Gemini. Er fungiert als zustandsbehafteter Konversationsmanager und multimodale Schnittstelle, die es Benutzern ermöglicht, Text- und Bild-Prompts an ein Sprachmodell zu senden und Antworten abzurufen.
Die Bibliothek nutzt ein Cookie-basiertes Authentifizierungssystem, das Session-Token aus dem lokalen Browserspeicher extrahiert, um Anfragen zu autorisieren. Zur Verwaltung von Zugriff und Konnektivität beinhaltet sie Proxy-basiertes Request-Routing, um regionale Beschränkungen zu umgehen und IP-Sperren zu vermeiden.
Das Projekt deckt Funktionen für multimodale KI-Analyse und die Pflege der Sitzungshistorie ab, um kontinuierliche Multi-Turn-Dialoge zu ermöglichen. Es bietet zudem Dienstprogramme zum Extrahieren von Bildlinks aus Antworten, zur Text-zu-Sprache-Konvertierung und zur automatischen Ausführung generierter Code-Snippets in einer lokalen Umgebung.
Features
- Gemini Integrations - Provides a Python wrapper for programmatically interacting with Google Gemini and Bard.
- Asynchronous Clients - Provides a non-blocking asynchronous Python interface for sending prompts to large language models and retrieving responses.
- Multimodal Input Processing - Processes combined text and image prompts to perform visual analysis through generative AI.
- Conversation State Management - Maintains session history and context to enable continuous, stateful multi-turn dialogues with the AI model.
- Conversation State Persistence - Maintains session objects and conversation history to support continuous multi-turn dialogues.
- Session-Cookie Wrappers - Uses a wrapper to extract browser session cookies for authenticating AI service requests without official API keys.
- Multimodal Analysis Tools - Provides utilities for analyzing combined text and image inputs to generate detailed descriptions of visual content.
- Visual Content Analysis - Analyzes images via text prompts to generate detailed descriptions of visual content.
- Stateful Session Management - Maintains conversation history and context within persistent objects to enable continuous dialogue.
- AI Prompt Sending APIs - Provides programmatic APIs to send text prompts to language models and retrieve structured responses.
- Browser Session Authentication - Uses extracted browser cookies to authenticate requests to AI services without official API keys.
- Browser Cookie Imports - Automatically retrieves session cookies from local browser storage to authorize API requests.
- Session Token Injection - Authenticates requests by extracting and using session tokens from local browser storage.
- API Wrapping - Wraps internal web service endpoints into a Python library to programmatically retrieve model responses.
- AI Content Workflow Automation - Integrates LLMs into scripts to automate code execution and text-to-speech synthesis.
- Code Execution Environments - Provides a mechanism to execute AI-generated code snippets within the local environment to produce results.
- Conversational Bot Development - Facilitates building chatbots that maintain stateful session history and context for multi-turn dialogues.
- Multimodal AI Toolkits - Ships a toolkit for performing integrated analysis of text and image data via generative AI services.
- Code Snippet Execution - Automatically runs code snippets received in AI responses within the local environment.
- Proxy Routing - Implements logic to route API requests through external proxy servers to bypass regional restrictions and IP blocks.
- External Proxy Forwarding - Forwards network traffic through external proxy servers to bypass regional restrictions and IP blocks.
- Asynchronous Request Execution - Implements a non-blocking asynchronous execution model for handling multiple network requests.
