Explore the platform

Understand how AI, Data, and Foundation work together in BullSequana AI.

Agentic Friendly

BullSequana AI separates customer-facing capabilities from the shared services that run them. Use this overview to find the right product domain before opening implementation detail.

Product domains

DomainUse it when you need toStart with
AIaccess models, ground responses in organizational knowledge, process documents, or build conversational and agentic applicationsExplore AI
Dataconnect, transform, govern, analyze, or operationalize enterprise dataExplore Data
Foundationoperate shared identity, inference, networking, observability, storage, databases, and delivery servicesExplore Foundation

These domains are responsibility boundaries, not isolated products. Most production solutions use all three.

Users reach AI and Data through the Portal, the BSQAI API, or the Developer Workspace and data services. The BSQAI API authenticates every call and checks the caller's permissions. Before a model call, it also checks the model's allow rules and the caller's monthly BCU caps, and it records the usage of every model call.

User entry points

BullSequana AI exposes different interfaces for different responsibilities:

  • Portal, with three areas:
    • OmniChat for chat, Build mode, files, Agents, Skills, Deep Research, Connectors, and API keys
    • Studio for models, experiments, deployments, the data catalog, data ingestion, and compute
    • Administration for authorized tenant and platform administration
  • BSQAI API for applications and developer tools
  • Developer Workspace for notebooks, pipelines, experiments, and model registration
  • Data services for integration, streaming, analytics, and dashboards
  • Platform CLI and GitOps for deployment and platform operation

The guided paths connect these interfaces into complete user journeys.

Choose the right starting point

Start in AI when the desired outcome is an AI experience: model access, RAG, document processing, voice, or agents.

Start in Data when the work begins with sources, pipelines, catalogs, notebooks, MLflow, streaming, or analytics.

Start in Foundation when the question concerns how the platform authenticates users, serves models, routes traffic, stores state, emits telemetry, or reconciles services.

If your goal is a task rather than a product-area explanation, use the guided paths. They connect the product domains in the order needed to reach an outcome.

From capability to implementation

The documentation uses three levels of detail:

  1. Guided paths describe an end-to-end customer outcome.
  2. Product pages explain the AI, Data, or Foundation responsibility involved.
  3. Reference pages document APIs, components, configuration, and operational behavior.

Guided paths are the primary entry point for customer work. Product pages add capability context, while component pages are reserved for deployment, architecture, and troubleshooting detail.

Common cross-domain flows

OutcomeAI contributionData contributionFoundation contribution
Grounded assistantmodel access, RAG, conversational experiencegoverned source data and ingestionidentity, inference, vector storage, observability
Document workflowextraction, classification, retrievalpipeline execution and artifact handlingcompute, storage, workflow services
Predictive applicationserving endpoints called through the BSQAI APIworkspace, experiments, pipelines, model registryserving runtime, security, telemetry
Governed analyticsnatural-language or embedded AI experiencecatalog, processing, streaming, BItenancy, databases, networking, audit

For a concise definition of tenants, teams, resources, and responsibility boundaries, read Core concepts.

Deployment control

BullSequana AI supports BullSequana hardware, private and sovereign cloud environments, and customer-managed Kubernetes infrastructure. The same product domains run across these environments while deployment guides capture environment-specific prerequisites and integrations.

Platform software, models, and documentation can be delivered through private channels for disconnected operation. Core services run inside the customer-controlled environment, and external telemetry is not mandatory.

Use Deploy and operate for the complete service lifecycle.

Typical user journey

  1. The user authenticates through the configured identity provider.
  2. BullSequana AI resolves the active organization, which is the user's tenant, with the user's groups and assigned permissions.
  3. The user opens the Portal, Developer Workspace, or another authorized service.
  4. Product interfaces expose only the models, data, applications, and operations allowed by that context. A model is listed and callable only when an allow rule lets the user use it.
  5. The BSQAI API checks the user's monthly BCU cap before a model call and records the call's usage.
  6. Foundation services provide inference, storage, workflows, and telemetry behind the experience.

On this page