Skip to main content

​
What is UtopikAI?

UtopikAI is a generative AI platform for your organization’s knowledge, tools, and teams. Connect workplace sources, choose the LLMs you need, and work through chat, assistants, agents, skills, tools, MCP servers, workflows, Studio, and tasks. Answers are grounded in the content you index. You pick LLM providers instead of being locked to one model. Native connectors and MCP servers link UtopikAI to the systems you already use. UtopikAI includes unified search across connected sources such as Slack, Google Drive, Confluence, GitHub, and Notion. Ask in natural language — for example whether a feature already exists, or where a policy or pull request lives — and get answers with citations from your knowledge. Authentication, roles, and chat history are built in for production use.

​
Présentation UtopikAI FR

​
Get started

  1. Open chat and try the built-in assistants — for example General, Art, and Search. You can use these as they are; you do not have to create your own first. Built-in agents may also be available depending on your tenant.
  2. Create a custom assistant for chat, or a custom agent, when you need something specific to your team.
  3. Attach tools, skills, and — for agents — MCP servers if they need to take actions.
  4. Use Studio to generate images, video, audio, and 3D.
  5. Use workflows for automations.
  6. Use tasks for scheduled or on-demand jobs.
  7. Optionally connect knowledge from Admin → Connectors so answers can be grounded in your sources. See Connector Overview.

​
Assistants and agents

Assistants are for talking with people in chat — questions, drafting, and guided help. Agents are for work that should run on its own, on a schedule, or inside a workflow. Agents also have a chat UI. They are two different entities; pick the one that matches the job. Create an assistant when people will talk to it in an interactive Q&A style. Create an agent when you also need autonomous or scheduled work. You can start with built-in assistants such as General, Art, or Search, or create your own. Each assistant and each agent has its own LLMs, knowledge, skills, tools, and prompts. Both use an eight-step setup: identity, access, knowledge and tools, Prompt Library (/ shortcuts in chat) and starter messages, advanced settings, space access, optional tasks, then publish.

​
Connectors

Connectors index your sources so assistants and agents can answer from your connected knowledge.
  • Hybrid search over connected content
  • Document sets to scope what an assistant or agent can search
  • Optional start dates, relevance filters, and multimodal indexing (text, images, tables) where configured
Administrators add and monitor connectors from Admin → Connectors. See Connector Overview.

​
Skills

Skills are reusable capabilities you install or create, then attach to a specific assistant or agent. Each assistant and each agent has its own skill list.
  • Install from the marketplace (name, space, visibility, version)
  • Create with details, instructions, access, then publish
  • Upload a skill.md or .zip package
See Skills Overview.

​
Tools

Tools let an assistant or agent fetch data, take actions, or generate content. Each assistant and each agent has its own tools. Manage connected apps in Admin → Tools (Connected Tools): See Tools Overview to connect an instance.

​
MCP Servers

MCP (Model Context Protocol) servers connect agents to external systems. Add them in Admin → MCP Servers, then attach a server to a specific agent.
  • List servers by name, URL, transport, auth, and connection status
  • Create a server in two steps: Server Details, then Auth & Transport
Each agent has its own MCP server list. Assistants do not share an agent’s MCP servers. See MCP Servers.

​
LLM

Each assistant and each agent has its own LLM attached.
  • Choose LLMs from providers such as GPT, Claude, Gemini, DeepSeek, Mistral, and Grok
  • Configure embedding models for knowledge search
See LLM and LLM Providers.

​
Workflows

Workflows are multi-step automations you create, run, and manage. They are not the same as Studio.
  • Create a blank workflow or start from a template
  • Build on a canvas with agents, tools, HTTP, and logic (Condition, While, User Approval)
  • Test in Draft, then switch to Production
See Workflows Overview.

​
Studio

Studio is where you generate images, video, audio, and 3D.
  • Image — text to image, or image to image
  • Video — text to video, image to video, or video enhancement
  • Audio — voiceovers and songs
  • 3D — text to 3D, or image to 3D
Each generator has a model picker, Advanced Mode, Target space, Generate, and My Creations. See Studio Overview.

​
Tasks

Tasks run assistant or agent work on a schedule or on demand, without a full workflow. Manage them from Admin → Task Management: You can also add a task in step 7 when creating an assistant or agent. Use tasks for recurring jobs such as daily summaries. Use workflows when the job needs multiple connected steps. See Tasks Overview.

​
Common use cases

  • Sales — customer and product facts during conversations
  • Marketing — brand and campaign materials with Studio images, video, audio, and 3D. See the Marketing use case.
  • Customer support — answers from docs and ticket history
  • Human resources — policies, onboarding, and internal knowledge
  • Technical teams — repositories, docs, and issue trackers
  • Operations — a single place to search, plus scheduled assistant or agent tasks
More examples: Use Cases Overview.