Deployment Models¶
Gaia is available in two distinct deployment models. Understanding the differences helps you choose the right approach and configure your application accordingly.
SaaS vs. Self-Managed¶
| Gaia SaaS | Gaia Self-Managed | |
|---|---|---|
| Hosting | Cohesity-managed cloud (AWS / Azure) | Your data center (Red Hat OpenShift) |
| Setup time | Minutes — create datasets and go | Requires GPU hardware, OpenShift, Gaia AI Engine install |
| Data residency | Data indexed in Cohesity cloud regions | Data stays entirely on-premises |
| LLM options | Cohesity Standard LLM, Cohesity Advanced LLM (Azure OpenAI GPT, Google Gemini 2.0) | Mistral 7B (runs locally on NVIDIA GPUs) |
| Updates | Automatic — Cohesity delivers new features | Manual — you control upgrade timing |
| Infrastructure | None to manage | OpenShift cluster + NVIDIA GPUs (L40S, H200, or RTX PRO 6000) |
| API | Same REST API | Same REST API (different base URL) |
| Best for | Fast adoption, cloud-first orgs | Regulated industries, data sovereignty requirements |
Same API, Different URL
Your application code works against both deployment models. The Gaia REST API is identical — only the base URL changes. SaaS uses https://helios.cohesity.com/v2/mcm/gaia, while Self-Managed uses your Helios Self-Managed FQDN.
Cloud Regions (SaaS)¶
| Cloud Provider | Regions |
|---|---|
| AWS | US-East-2 (Ohio), US-West-2 (Oregon), Canada (Central) |
| Microsoft Azure | Central US (Iowa) |
When creating datasets, you select a region for storage and processing. Choose the region closest to your data and users.
LLM Options¶
SaaS¶
| Model | Powered By | Best For |
|---|---|---|
| Cohesity Standard LLM | Azure OpenAI GPT or Google Gemini 2.0 | General-purpose queries, most use cases |
| Cohesity Advanced LLM | Azure OpenAI GPT or Google Gemini 2.0 | Complex analysis, nuanced answers |
You can set your preferred LLM in Settings > Preferences. If your account has both entitlements, you can switch between them.
Self-Managed¶
| Model | Details |
|---|---|
| Mistral 7B | Runs locally on your NVIDIA GPUs. No data leaves your network. |
Specifying LLMs via API
When calling POST /ask, you can pass llmName to select a specific LLM. If omitted, the account default is used. Use GET /llms to list available models.
Supported Object Types¶
| Object Type | SaaS | Self-Managed |
|---|---|---|
| Microsoft 365 Mailbox | ||
| Microsoft 365 OneDrive | ||
| Microsoft 365 SharePoint Site | ||
| Isilon NAS | ||
| NetApp ONTAP | ||
| Generic NAS | ||
| Physical Servers | ||
| Views |
Supported File Types¶
Gaia can index the following file formats (max 100 MB per file):
.doc .docx .xls .xlsx .ppt .pptx .pdf .odf .rtf .txt .html .xml
Access Roles¶
| Role | Capabilities |
|---|---|
| Gaia Admin | Create/edit/delete datasets, manage users, chat, select datasets, recover documents, full API access |
| Gaia Viewer | Chat with assigned datasets, view Topic Explorer, provide feedback. Cannot create datasets or access other Cohesity apps. |
For App Developers
When building applications, your backend typically authenticates with an API key that has Gaia Admin privileges. End users interact through your app's session-based auth — they never need direct Gaia access.
Next Steps¶
- Architecture Overview — How the pieces fit together.
- Prerequisites — Set up your development environment.
- Authentication — Connect to the Gaia API.