AI · Database · Assistant
Debi
AI-Powered DBA Assistant
One unified API endpoint. Four LLM providers. Switch between Claude, GPT-4, Gemini, or NVIDIA NIM with a single config change. Zero code rewrites.
Provider Routing
One request. Any model.
Set LLM_PROVIDER=anthropic and all requests go to Claude. Change it to google and it routes to Gemini. No deployments, no code.
Built for developers.
Provider Agnostic
Switch between Claude, GPT-4o, Gemini, or NVIDIA NIM by changing one environment variable. Zero code changes.
Unified Endpoint
Single POST /llm/generate endpoint. Your application calls one URL regardless of which LLM is running under the hood.
Pydantic Validated
Strict request/response validation with Pydantic models. Type-safe from input to output with clear error messages.
Cloud-Ready
PostgreSQL connection pooling pre-configured for AWS, Azure, GCP, or local databases. Deploy anywhere.
Cloud Ready
Deploy anywhere.
PostgreSQL connection pooling pre-configured for all major cloud platforms. Docker included.
One endpoint for all your LLMs.
Switch providers without touching your application code.