Does Enterprise h2oGPTe cite its sources?

Yes. Answers cite the document and passage they came from. Multi-step retrieval and reasoning are used when a single lookup is not enough.

Which models can I use with Enterprise h2oGPTe?

OpenAI, Anthropic, Amazon Bedrock, Azure OpenAI, Databricks, self-hosted vLLM endpoints, models fine-tuned in H2O Cascade, and open-weight Danube and Mississippi. Intelligent model routing selects per step; you set the policy.

Is the Super Agent a separate product from Enterprise h2oGPTe?

No. The H2O AI Super Agent is the Agentic AI orchestration layer that ships inside Enterprise h2oGPTe. Buying the product gives you the core.

Can Enterprise h2oGPTe run air-gapped?

Yes, on H2O AI Hybrid Cloud with internal model endpoints. Documents, embeddings, and fine-tuned weights stay in your environment.

Enterprise h2oGPTe

The enterprise agentic and generative AI platform: Agentic RAG with citations, Document AI, any model from any host, and the H2O AI Super Agent, in one governed product.

Request a demo Documentation

The record

9agent types in the Super Agent

27built-in tools, plus MCP servers

35%of the prior AI cost at AT&T, at 91% accuracy

2023h2oGPT paper, arXiv 2306.08161

Key takeaways

Enterprise h2oGPTe is H2O.ai's enterprise GenAI and agentic platform: chat and Agentic RAG with citations, Document AI, multi-model support, and the home of the H2O AI Super Agent.
It connects to OpenAI, Anthropic, Bedrock, Azure OpenAI, Databricks, vLLM, and Cascade-tuned models, with cost controls per user, team, and model.
It ships the H2O AI Super Agent with nine agent types, 27 tools, MCP servers, and A2A endpoints.
H2O.ai fine-tuned small language models reached 91% accuracy at 35% of the prior cost at AT&T; the Super Agent runs inside Ask AT&T.

What it does

Three capabilities in one tool, governed as one.

Enterprise GenAI, H2O Document AI, and the H2O AI Super Agent share one platform, one identity model, one audit trail, and one bill.

Chat and Agentic RAGAsk across contracts, policies, tickets, wikis, and scans. Answers cite the document and passage, with multi-step retrieval when one lookup will not do.
Document AIIngest PDFs, scans, faxes, images, and forms. OCR and layout understanding on Mississippi vision-language models. Tables and fields extracted and queryable.
Any model, any hostConnect OpenAI, Anthropic, Bedrock, Azure OpenAI, Databricks, or your own vLLM. Intelligent routing, with cost controls per user, team, and model.
The H2O AI Super AgentPlain English in; a plan, the tool calls, and the result out. Nine agent types, including Deep Research, Data Science, Tool Builder, and Code.
Make and connect agentsAgents Builder and Flow for custom agents. MCP servers for tools, local or remote. A2A endpoints so your agents talk to others.
GovernSSO and RBAC, collections with access control, and cost per user. Every prompt, retrieval, tool call, and answer logged, and evaluated in Eval Studio.

How you use it

From a folder of documents to a governed assistant, to an agent that acts.

01IngestUpload or connect document sources. Document AI parses the scans, tables, and forms for you.
02AskChat with citations to the source. Choose the model per collection, and set cost limits.
03AutomateTurn a workflow into an agent. Connect tools over MCP, and add an approval gate.
04GovernAudit every step, evaluate the RAG and agents in Eval Studio, and attribute the cost.

Where it fits

The consumer side of the generative lane, and the home of the Super Agent.

Cascade produces models; h2oGPTe consumes them alongside external LLMs and turns knowledge into answers and actions. The Super Agent lives here. Highlighted are h2oGPTe and the Super Agent.

Details

What ships today

RetrievalAgentic RAG with citations; collections; multi-step retrieval and reasoning

DocumentsPDF, Office, HTML, images, scans, faxes; OCR and layout via Mississippi VLMs; tables and fields

ModelsOpenAI, Anthropic, Bedrock, Azure OpenAI, Databricks, vLLM, self-hosted; Cascade-tuned models; intelligent routing

AgentsH2O AI Super Agent: General, Deep Research, Task, Data Science, Web Search, Tool Builder, Agents Builder, MCP Runner, Code

Tools27 built-in, Python and shell, 16 browser actions; MCP servers (local and remote); A2A protocol

AccessSSO/RBAC, collection-level permissions, per-user and per-team cost controls

APIsREST API and Python client

DeploymentOn-premises, air-gapped, private cloud, or cloud. Air-gapped uses internal model endpoints

With the Super Agent

This is where the orchestration layer runs.

Enterprise h2oGPTe is the product. The H2O AI Super Agent is the orchestration layer inside it: it makes, orchestrates, and manages agents across predictive, generative, agentic, and observability. Buy the product, get the core.

One agent, plus nine, plus yoursOne general-purpose agent, nine specialized types, and the custom ones you build yourself.

Your agents, hosted and governedBring LangGraph, CrewAI, or the Agents SDK. They run alongside the Super Agent, hosted and governed.

Grounded in your documentsAgentic RAG is how every plan starts, so the agent reasons over your knowledge instead of guessing.

Run it anywhere

You operate it, or H2O AI operates it for you.

Same platform either way: the same capabilities, the same governance, and the same audit trail. The choice stays open, so you can switch between them as the business changes, without rebuilding what runs on top.

We run itH2O AI Managed Cloud (HAMC)A dedicated, single-tenant cloud that H2O AI provisions, operates, and secures, across fault-tolerant availability zones in the region you choose. Your data never leaves it. Zero-trust access, encrypted in transit and at rest, constantly monitored.AWSAzureGoogle CloudSOC 2 Type 2 · HIPAA / HITECHLearn more You run itH2O AI Hybrid Cloud (HAHC)The full H2O Sovereign AI Platform, entirely within your control, on your own infrastructure or fully air-gapped. You own the environment, the updates, and the compliance posture. Installed and upgraded with Terraform, Helm, or Replicated.Your cloud accountYour data centerAir-gappedLearn more

H2O AI Cloud for Government

The H2O AI Cloud for Government has achieved FedRAMP High Certification, delivering the highest standard of cloud security so federal agencies can harness predictive and generative AI without compromising data sovereignty or compliance.

Learn more View on marketplace

Works with

The tools the Super Agent runs.

H2O CascadeFrontier capability, flowing down into a model you own.Explore H2O Eval StudioSystematic evaluation, with a human calibrating the bar.Explore

Videos

Watch Enterprise h2oGPTe in action

VideoBuilding, Evaluating, and Optimizing your RAG App for Production | Jerry LiuWatch

Videoh2oGPT DemoWatch

FAQ

Questions people ask before they call us.

Enterprise h2oGPTe is H2O.ai's enterprise generative AI and agentic platform. It provides chat and Agentic RAG over your documents with citations, Document AI for scans and forms, multi-model support with cost controls, and contains the H2O AI Super Agent with nine agent types, MCP, and A2A.

Next step

Bring a folder of your ugliest documents.

We will ingest them, answer your questions with citations, and turn one workflow into an agent, in a single session.

Request a demo Talk to H2O