Develop
Developer entry points for models, voice, document AI, data connectors, developer tools, and application delivery on BullSequana AI.
Development on BullSequana AI follows six practical tracks:
- calling generative models through the BSQAI API, with the Responses, Chat Completions, Anthropic Messages, embeddings and rerank APIs, and calling predictive models through their serving endpoints
- building voice conversations
- building document AI and retrieval flows
- building data connectors and operating data or ML pipelines
- connecting coding agents and other developer tools
- deploying applications on the platform
This section is for developers, solution teams, and platform teams building on BullSequana AI.
Recommended background
Helpful experience for this section includes:
- general software engineering and API integration
- Docker and local development environments
- bearer-token or API-key based authentication
- Kubernetes basics
- Helm packaging
- GitOps concepts when deploying through
Argo CD
The more detailed pages in this section add more specific guidance where needed.
Start here
Use local models via API
Authenticate, discover models, and call the Responses, Chat Completions, and embeddings APIs.
Use predictive models via API
Send predictions to classic machine learning models deployed from the model registry.
Build voice conversations
Stream speech, handle interruptions, and retain conversation history.
Build and register a connector
Package customer-owned connectors for managed data ingestion, including Meltano/Singer adapters.
Developer tools
Connect Claude Code, Codex CLI, OpenCode, Continue, and GitHub Copilot to platform models.
Deploy apps on the platform
Package and deliver applications with Argo CD, Helm, and Keycloak SSO.