AI & Agents

Effortlessly switch between AI providers or use multiple models simultaneously in the same workflow. No code changes, no integration headaches—just configure and deploy.

As new AI providers and models emerge, appRules.AI rapidly integrates them into the platform. Your workflows stay current without rewrites or migrations.

Leverage the most cost-effective AI provider for each task. Route simple queries to efficient SLMs and complex analysis to powerful LLMs—automatically balancing performance and budget.

Deploy multiple AI providers simultaneously for critical decisions. Compare outputs, validate results, and ensure accuracy across mission-critical workflows.

Select the optimal AI model for each task and optionally associate it with a Hub workflow activity for enhanced coordination and orchestration.

Configure queries for the agent's process by simply selecting the data source and related entities to use—no complex programming required.

Define exactly where and how results are stored by selecting the related fields for agent outputs—ensuring seamless integration with your existing systems.

Support for optional actions and conditions gives you precise control over when and how the agents execute within your workflow.

AI agents include expertly-crafted starting prompts for each process. Optionally modify and customize them for your specific use case—start with excellence, refine to perfection.

Deploy sophisticated data classification with single or multi-label assignment, confidence scoring, and flexible output configuration—transforming chaotic data into structured intelligence.

Craft precision AI prompts with dynamic variables, conditional logic, and reusable templates. The SmartPrompt Builder is the intelligence layer of every custom agent you build — start with an expert template and tailor it to your exact use case.

Assemble your custom AI Agent visually using the drag-and-drop Workflow Designer. Chain AI decisions, data operations, and integrations into a complete end-to-end agent — no programming required.

Hundreds of built-in Actions and Conditions give your custom agents the power to interact with data sources, files, APIs, and enterprise systems with precision — all configurable without code.

Every custom agent you build inherits appRules' GuardRailed execution model — AI advises, appRules acts. Your agents are safe, auditable, and fully under enterprise control from day one.

Specify any input folder and the agent processes every file within it — one by one, at scale — with no manual intervention required.

Choose what to send to the AI: file attributes for quick triage, file contents for deep reading, or a raw file upload for vision and multimodal analysis.

Start with an expertly crafted prompt template and use the SmartPrompt Builder to tailor it to your exact use case. The AI is instructed to return a clean, structured decision — "Delete", "Move", or a classification label — so the appRules workflow always knows exactly what to do.

The agent never modifies, moves, or deletes files directly. It returns a decision; the appRules workflow activity executes the action. This separation of concerns keeps every operation safe, traceable, and fully under your control.

Define your own classification labels and map each one to a folder, action, or workflow. Files are routed to the right destination based on AI judgment — no hardcoded rules required.

Each classification carries its own result action. Copy files to a local subfolder while keeping the original in place, move them to a destination folder, or transfer them remotely via FTP or other protocols — configured independently per classification so every file goes exactly where it needs to go.

Optionally save the AI's reasoning alongside every processed file. When enabled, the appRules AI File Agent creates a companion .ai.notes.txt file in the destination folder — capturing the AI's classification rationale in plain text. Every decision becomes auditable, every action explainable. Ideal for resume screening, academic paper review, legal document triage, and any workflow where understanding why a file was classified matters as much as the classification itself.

From resume screening and invoice sorting to legal document triage and support ticket routing — any folder filled with documents that need intelligent classification is a perfect fit for the appRules AI File Agent.

Specify a parent folder and the AI Folder Agent processes every subfolder within it — treating each as a complete document package. The AI sees the full picture before making its classification decision.

File attributes, metadata, and contents from all files within each subfolder are aggregated and sent to the AI together — enabling richer, more accurate classification decisions than any single file could provide.

The AI returns a single classification for the entire subfolder. appRules then copies, moves, or deletes the complete package — keeping every related file together throughout the workflow.

Optionally save the AI's reasoning alongside every processed subfolder as a companion .ai.notes.txt file — capturing the classification rationale for the entire package in plain text. Every decision auditable, every action explainable.

The AI advises, appRules acts. The AI Folder Agent never directly modifies, moves, or deletes folders. It returns a classification; the appRules workflow activity executes the action — safe, traceable, and fully under enterprise control.

Connect to 14+ vector databases including Azure AI Search, Pinecone, Weaviate, Qdrant, Milvus, and more. Switch providers instantly or implement hybrid strategies without code changes.

Generate high-quality embeddings using any AI provider or specialized services like HuggingFace. Optimize for accuracy, speed, or cost based on your specific use case.

Dramatically improve search precision with AI-powered reranking services. Combine with metadata filters for hybrid retrieval that outperforms pure semantic search.

Extend knowledge discovery beyond your documents with Bing, Brave, and Google web search integration. Ground AI responses in current information and external sources.

Process DOCX, DOC, PDF, HTML, RTF, ODT, and TXT files seamlessly—including complex documents with tables, images, and formatting

Simple, Fixed Size, By Paragraphs, By Sentences, Semantic Sections, By Token Count, Recursive, Custom and Embedded Extract—each optimized for specific scenarios

Fine-tune max/min chunk sizes, overlap percentages, paragraph separators, and sentence delimiters for optimal results

No programming, no iterations, no complexity—just select, configure, and execute through our visual interface

Works out of the box with any MCP-compatible AI client. Claude Desktop, Cursor, Windsurf, and future clients connect with a single configuration entry — no custom integration required per tool.

Ask your AI assistant to run an appRules workflow in plain English. The MCP server translates the intent into a precise API call to your appRules Web API — instantly, without writing code.

Credentials stay in your environment variables — never exposed in prompts or tool calls. appRules handles all authentication, so your security posture is unchanged.

Run workflows by Job ID for quick named executions, or pass full project credentials with runtime arguments for dynamic, parameterised workflow runs — matching exactly how your teams already work.

The MCP Server calls the same appRules Web API already used by your CLI tools. Every existing workflow, rule, and integration works immediately — nothing needs to be rebuilt or reconfigured.