casibase/casibase

Features

• Retrieval-Augmented Generation - Augments LLM responses by fetching relevant context from a stored knowledge base before generating an answer.
• LLM Application Orchestrators - Orchestrates multi-turn conversations with large language models and manages RAG knowledge bases.
• Vector Knowledge Bases - Ingests documents and converts them into vector embeddings for efficient similarity search and retrieval.
• Knowledge Base Retrieval - Converts both knowledge base content and user queries into vectors to find the most relevant matches quickly.
• Multi-turn Interaction Managers - Manages stateful multi-turn dialogues with large language models, maintaining context across exchanges.
• Model Provider Integrations - Connects to a wide range of language and embedding models through a unified interface.
• OpenAI-Compatible APIs - Exposes an OpenAI-compatible API endpoint so existing clients can use the assistant as a backend.
• Provider Switching - Switches between major LLM providers per conversation without requiring code changes.
• Knowledge Store Configurations - Connects storage, embedding, and AI model providers into a single store that powers knowledge-base queries and chatbot interactions.
• Model Selection Policies - Chooses among supported language models from providers like OpenAI, Azure, and HuggingFace.
• AI Model Integrations - Connects to cloud-hosted or open-source AI models and lets users switch between them for dialogue generation.
• Autonomous Agent Loops - Ships an autonomous agent loop that drives a browser, executes commands, and integrates MCP servers.
• Autonomous Task Execution - Runs autonomous agent loops that decompose goals into sub-tasks, invoke tools, and iterate until completion.
• Unified Provider Interfaces - Routes requests to any of over 30 supported AI services through a single unified interface.
• Context-Aware Conversation Managers - Combines a language model with retrieved domain knowledge to produce contextually relevant answers.
• Conversational AI Assistants - Orchestrates multi-turn dialogues with an AI assistant, maintaining context and storing history.
• LLM Service Connectors - Integrates with external LLM services to generate responses and power conversational AI features.
• AI Chat Interfaces - Provides a web-based chat interface for sending messages to connected language models.
• LLM Orchestration - Orchestrates multi-turn dialogues by routing user messages through a large language model.
• Service Provider Abstractions - Routes requests through a unified interface that dynamically loads provider implementations for LLMs, embeddings, and storage.
• Multi-Model AI Orchestrators - Switches between over 30 model providers per conversation without code changes.
• Text Chunks - Splits long documents into smaller segments to fit within model context windows.
• RAG Knowledge Management - Ingests documents, generates vector embeddings, and retrieves context to augment AI responses.
• Vector Embeddings - Converts documents into dense vector representations through configurable chunking, embedding, and indexing stages.
• Cloud Embedding Provider Integrations - Integrates with external embedding services to convert text into vector representations for semantic search.
• Natural Language Querying - Performs similarity search on stored vectors and constructs prompts for a language model to answer domain-specific questions.
• Automatic Vector Embeddings - Converts uploaded files into dense vector representations automatically when files are added to a store, processing them in the background.
• Conversation History Stores - Stores and retrieves chat history with full message content and metadata for review.
• Semantic Knowledge Base Search - Finds relevant information by comparing vector representations of user questions against stored knowledge vectors.
• Automatic Billing Records - Automatically creates billing records after each AI response, keeping billing synchronized with usage.
• Embedding Generation - Automatically splits files into chunks, generates embeddings, and stores them for retrieval.
• Vector Data Conversion - Transforms text and other data into dense vector representations for efficient storage, retrieval, and analysis.
• Embedding Provider Configurations - Links a vector embedding service to the system for converting text into numerical representations used in search and retrieval.
• Chat Session Management - Lists and manages chat sessions with details like name, user, token count, and cost.
• Role-Based Access Controls - Enforces permission checks at router, controller, and service layers to isolate data and restrict operations by user role.
• Private Knowledge Base Imports - Uploads documents in various formats and converts them into vector embeddings for semantic search and retrieval.
• Role-Based Access Control - Checks a user's assigned role on each request and blocks actions that the role is not permitted to perform.
• Single Sign-On - Delegates identity verification to an external SSO provider so users log in once and access the application without separate credentials.
• Third-Party Authentication Providers - Logs in using accounts from external services like GitHub, Google, or WeChat, with support for plugin-based extensions.
• AI Usage - Tracks token consumption and cost per provider, model, and user, with real-time activity monitoring, success/error rates, and interactive charts.
• AI Cost Monitoring - Records token consumption and cost per AI message as structured transactions for traceability.
• Token Usage Analytics - Counts tokens consumed per conversation and reports totals to track LLM spending.
• Cost and Token Trackers - Records token consumption and pricing per message automatically for per-conversation billing.
• Multi-Store Isolations - Maintains separate knowledge stores with independent model, storage, and embedding configurations for different user groups.
• AI Conversation Managers - Manages interactive question-answer sessions where user queries are processed by AI models.
• Web Search & Fetch Configurations - Toggles web search during chat to fetch current information with source URLs.
• Answer Source Tracings - Shows knowledge fragments and web sources with relevance scores for answer verification.
• Conversation Session Initializers - Opens a chat interface on login to immediately begin a dialogue with an AI assistant.
• Collapsible Reasoning Displays - Displays the model's step-by-step reasoning in expandable sections for user inspection.
• Third-Party Model Integration - Integrates third-party language model services to power conversational and reasoning tasks.
• Browser-Based Remote Desktops - Provides browser-based RDP access to manage nodes and execute commands on remote machines.
• Event-Driven Workflow Orchestrators - Orchestrates multi-step workflows with conditional branching and parallel execution through a visual drag-and-drop editor.
• Multi-Store Management - Creates and activates different stores with their own model, storage, and embedding services.
• Store Activation Modes - Activates a mode where users can select from multiple configured stores with their own models.
• Pre-Processing Fund Validations - Checks estimated token cost against user balance via dry-run before processing requests.
• Store Selectors - Lets users choose a specific store from a dropdown when starting a new chat.
• External Storage Connectors - Connects to cloud or local storage backends to persist and load files and knowledge artifacts.
• Charge Event Logging - Logs each charge event as a transaction record for auditing and reconciliation.
• Cross-Store Vector Search - Lets a store query vectors from other selected stores, creating a centralized knowledge hub without duplicating content.
• Multi-Backend Storage Management - Saves and retrieves data through a unified interface across local storage or cloud services like AWS S3 and Azure Blob.
• Usage Billing Systems - Stores transaction records in a shared billing system to unify AI usage costs with other expenses.
• Application Lifecycle Management - Controls the full lifecycle of deployed applications, from creation through termination, with namespace isolation for security.
• Deployment Environments - Uses Kustomize to maintain consistent base configurations while allowing environment-specific overrides for replicas, images, and limits.
• Cloud Storage Integrations - Connects to cloud storage services like AWS S3 and Azure Blob for file persistence.
• Container Orchestration Management - Creates, deploys, and manages containerized workloads on a Kubernetes-based platform.
• Containerized Infrastructure Managers - Provides a unified interface for connecting to Docker hosts and Kubernetes clusters to manage container lifecycles.
• Kubernetes Application Deployments - Deploys and manages containerized applications on Kubernetes using reusable blueprint templates.
• Blueprint-Based Deployments - Deploys containerized applications by instantiating reusable Kubernetes templates with environment-specific overlays.
• Kustomize Deployment - Applies user-provided settings like replica count or domain name as overlays on a base template using Kustomize.
• Remote Management Gateways - Provides a web-based gateway for browser access to remote nodes via SSH, RDP, and VNC.
• Storage Backend Integrations - Integrates with external storage backends to persist files, documents, and other binary data.
• Storage Connection APIs - Attaches a remote storage service to the system so files and data can be saved and retrieved from outside the local environment.
• Template-Based Deployment - Deploys containerized applications by selecting pre-configured Kubernetes templates with parameter customization.
• Chat History Viewers - Displays a searchable log of all messages with metadata like timestamps and token usage.
• Multi-Protocol Remote Session Launchers - Establishes secure remote connections to machines and tracks session activity with audit logs.
• Messaging Platform Integrations - Integrates with enterprise chat tools to receive messages, generate AI replies, and stream responses in real time.
• Multi-Protocol Remote Node Connectors - Establishes remote connections to managed nodes using SSH, RDP, VNC, or Telnet protocols for administration or desktop access.
• Remote Access Gateways - Establishes browser-based connections to remote nodes via SSH, RDP, and VNC through a Guacamole proxy.
• Local Language Model Hosting - Hosts open-source or custom models on the user's own infrastructure using an OpenAI-compatible API.
• Custom AI Provider Registration - Registers external AI model services so the assistant can use them for chat, analysis, and question answering.
• Citation-to-Source Navigators - Clicks a referenced file in a chat response to open the store's file viewer with that document selected.
• Visual Workflow Builders - Composes multi-step pipelines with a drag-and-drop editor supporting conditional branching and parallel execution.
• Application Frameworks - RAG knowledge database with web UI and SSO.
• Data Storage Systems - Acts as a RAG-focused knowledge database with SSO.

Star history