What is MINDTRICKS AI?

At a glance

The product is built around a few ideas: you control which capabilities are on, the system plans before it acts, and heavier outputs (like long documents or diagrams) are handled in dedicated steps instead of cluttering every chat reply.

• Chat-first experience with streaming answers and multi-turn context
• Optional web and knowledge-base search when a question needs fresh or private sources
• Retrieval augmented generation (RAG): your uploaded material is indexed, retrieved, and passed to the model for grounded responses
• Agentic orchestration so each message follows a sensible plan instead of one fixed pipeline
• Image generation and editing tools for visual workflows alongside text
• Voice and Canvas-style collaboration where enabled—see AI Product Policy for the formal service description

Knowledge pipeline: indexing → retrieval → generation

For answers that use your documents, the flow is: prepare sources, find the best excerpts, then generate a response that cites that context. These docs walk through each stage:

• Document indexing — chunking, embeddings, and storage
• Document retrieval — similarity search and ranking
• AI response generation — models, prompts, and streaming output

How the assistant plans and responds

The assistant can branch: plan what to do, gather evidence when needed, write the main reply, then optionally run follow-up jobs (for example diagrams or Canvas documents) only when relevant and allowed.

Read the agentic overview for the full workflow, including how tool choices and LangGraph-style orchestration stay behind the scenes.

Image product

Separate documentation covers generative image creation and post-processing utilities:

• Generative image generation
• Image post-processing tools

Privacy and terms

Data handling, retention, and your rights are described in the Privacy Policy. Usage rules, disclaimers, and service scope are in the AI Product Policy & Terms of Service.