Put your organisation's knowledge to work

Information overload

Employees lose hours every week looking for information they know exists somewhere. New hires take months to get up to speed, and expertise walks out the door when people leave.

"Shadow AI"

Staff paste contracts, customer data and internal reports into public chatbots to work faster. Productivity goes up, but so does the data-protection risk, and IT can't see any of it.

Generic AI doesn't know you

Public AI has never read your HR policy, product catalogue or last year's tender, so it gives confident answers that are wrong.

Answers you can trust

• Cited answers. Every response shows which documents it came from.
• Query mode answers only from your documents, and says so when the answer isn't there. Chat mode blends in general AI knowledge for drafting and brainstorming.
• Accuracy controls: relevance thresholds, re-ranking for more precise search, and document pinning, so the AI reads critical documents in full.
• Automatic sync re-reads web pages and connected sources when they change.

Workspaces for every team

• Dedicated workspaces such as "HR Policies", "IT Helpdesk", "Tender Library" or "Client X", each with its own documents, AI model, instructions and members.
• Conversation threads keep separate topics organised.
• Custom instructions, with variables for the user's name, today's date and company details, set the right tone and rules for each workspace.

AI agents & no-code automation

• Built-in skills: document search, web search and browsing, website scraping, summarisation, chart generation, file creation (Word, Excel, PDF, PowerPoint) and a SQL agent that answers questions from your databases.
• No-code Agent Flows: business users build multi-step automations in a visual editor, without writing code.
• Custom skills for anything bespoke, built by developers.
• Human in the loop: agents can stop to ask clarifying questions before acting.

Connects to your systems

• Model Context Protocol (MCP) support connects agents to a fast-growing ecosystem of tools and to your internal systems.
• Identity-aware integrations pass the user and workspace to your systems, so agents only see what each person is allowed to see.
• Intelligent tool selection gives agents only the tools each request needs, keeping responses fast and costs low.
• Full REST API with interactive documentation, to embed Mindspace in your portals and apps.

Any AI model, no lock-in

• 35+ AI providers: OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini, AWS Bedrock, Mistral, DeepSeek, Groq and more.
• Private, local models, so sensitive data never leaves your environment.
• Model Router sends each question to the best-fit model: low-cost for simple questions, most capable for complex reasoning, local for sensitive topics.
• Switch providers as the AI market evolves, without rebuilding your knowledge base.

Enterprise-grade security & governance

• Role-based access: Admins run the system, Managers look after workspaces and users, and everyone else only sees the workspaces they're added to.
• Single sign-on with your identity provider, delivered by Hiend.
• Event and chat logs for audit and quality review, exportable to CSV and JSON.
• Right-to-erasure: one deletion removes a document and its embeddings everywhere, helping you meet PDPA obligations.

Your brand, everywhere

• White-labelling: your name, logo, colours, favicon, footer links and support email.
• Website & intranet chat widget that answers from an approved workspace, with domain restrictions and daily chat limits.

Personal & productive

• Memories: users can opt in to have the assistant remember their role, preferences and projects. Each user's memories are private to them.
• Multimodal: analyse images and screenshots, generate images, and listen to answers read aloud.
• Export conversations for record-keeping or fine-tuning.